LETTERS
PUBLISHED ONLINE: 12 AUGUST 2012 | DOI: 10.1038/NNANO.2012.128
Printing colour at the optical diffraction limit
Karthik Kumar1†, Huigao Duan1†, Ravi S. Hegde2, Samuel C. W. Koh1, Jennifer N. Wei1
and Joel K. W. Yang1 *
The highest possible resolution for printed colour images is
determined by the diffraction limit of visible light. To achieve
this limit, individual colour elements (or pixels) with a pitch of
250 nm are required, translating into printed images at a resolution of ∼100,000 dots per inch (d.p.i.). However, methods for
dispensing multiple colourants or fabricating structural colour
through plasmonic structures have insufficient resolution and
limited scalability1–6. Here, we present a non-colourant method
that achieves bright-field colour prints with resolutions up to
the optical diffraction limit. Colour information is encoded in
the dimensional parameters of metal nanostructures, so that
tuning their plasmon resonance determines the colours of the
individual pixels. Our colour-mapping strategy produces
images with both sharp colour changes and fine tonal variations,
is amenable to large-volume colour printing via nanoimprint
lithography7,8, and could be useful in making microimages for
security, steganography9, nanoscale optical filters6,10–12 and
high-density spectrally encoded optical data storage.
Abbe’s classical diffraction limit13 states that the minimum
resolvable distance between two closely spaced objects is at best
half the wavelength used for imaging. Hence, assuming 500 nm as
the mid-spectrum wavelength for visible light, this limit dictates
that an idealized lens-based optical microscope can resolve juxtaposed colour elements down to a pitch of 250 nm. However,
colours produced by depositing materials, such as dyes and
quantum emitters2,3, or by iridescence of periodic structures1,6,8,14
cannot yet achieve this printing resolution. Industrial techniques
such as inkjet and laserjet methods print at sub-10,000 d.p.i. resolutions because of their micrometre-sized ink spots. Research-grade
methods15,16 are capable of dispensing dyes at higher resolution but
are serial in nature and, to date, only monochrome images have
been demonstrated.
Plasmon resonances in metal films have been used in
macroscopic colour holograms5, full colour filters and
polarizers6,10–12,17,18. The colour filters in particular exhibit the
phenomenon of extraordinary optical transmission (EOT, an
effect of Fano resonance19,20) through periodic subwavelength
holes in the film21–24. The colours produced are set by the periodicity
of the structures, so multiple repeat units are required, resulting in
large (micrometre-sized) pixels10,11. In an alternative arrangement,
small (tens of nanometres) isolated metal nanoparticles can be
used; these scatter colours depending on their shapes and sizes,
but do not scatter strongly enough to be viewed plainly in a
bright-field reflection microscope, especially when deposited in
direct contact with a substrate4. Hence, the challenge still remains
to create pixels that support individual colours, can be miniaturized
and juxtaposed at the optical diffraction limit, and also produce
vivid colours when observed in a bright-field optical microscope.
Here, we increase the scattering strength of particle resonators by
raising them above a metal backreflector to obtain 250 nm-pitch
pixels that reflect individual colours without a dependence on
periodicity. A schematic representation of two such pixels is provided in Fig. 1a, where each pixel consists (arbitrarily) of four nanodisks that support particle resonances. These disks are raised above
equally sized holes on a backreflector. Crucially, this backreflector
plane functions as a mirror to increase the scattering intensity of
the disks (Supplementary Fig. S1). A key feature of such structures
is their ease of fabrication and throughput scale-up by means of
nanoimprint lithography (NIL); indeed, similar structures have
been investigated previously for other applications22,23,25–27.
We introduce the concept that small groups of these structures
(with different diameters D and gaps g) reflect different colours
a
g
Nanodisk
(Ag\Au)
Nanoposts
(HSQ)
b (i)
D
Backreflector (Ag\Au)
Si substrate
(iii)
(iv)
140 nm
(ii)
Nanodisk Backreflector
Figure 1 | Working principle and fabrication process for high-resolution
plasmonic colour printing. a, Interaction of white light with two closely
spaced pixels, each consisting of four nanodisks. As a result of the different
diameters (D) and separations (g) of the nanodisks within each pixel,
different wavelengths of light are preferentially reflected back. b, Method of
fabrication of nanostructures. (i) A 95-nm-thick layer of HSQ is spin-coated
onto a silicon wafer piece and patterned using EBL. (ii) The unexposed
portions of the HSQ are developed away using a salty developer (see
Methods), leaving HSQ nanoposts. (iii) The nanoposts and backreflector are
coated using a single metal evaporation step. (iv) A 708 side-angle SEM
image of nanostructures after metal deposition. Colour information is
encoded in the nanopost diameter and spacing of the resist structure.
1
Institute of Materials Research and Engineering, A*STAR, 3 Research Link, Singapore 117602, 2 Institute of High Performance Computing, A*STAR,
1 Fusionopolis Way, Singapore 138632; † These authors contributed equally to this work. *e-mail: [email protected]
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology
© 2012 Macmillan Publishers Limited. All rights reserved.
1
LETTERS
NATURE NANOTECHNOLOGY
a
DOI: 10.1038/NNANO.2012.128
D (50–140 nm)
b
12 µm
g (30–120 nm)
c
(ii) Simulations
(i) Experiments
140 nm
140 nm
130 nm
130 nm
d 800
Dip (experiments)
120 nm
120 nm
Dip (simulations)
Peak (experiments)
110 nm
110 nm
Peak (simulations)
80 nm
70 nm
60 nm
50 nm
100 nm
90 nm
80 nm
Wavelength (nm)
90 nm
Reflectance (a.u.)
Reflectance (a.u.)
700
100 nm
600
70 nm
500
60 nm
50 nm
400
400 500 600 700 800
Wavelength (nm)
400 500 600 700 800
Wavelength (nm)
50
60
70
80 90 100 110 120 130 140
Diameter (nm)
Figure 2 | Optical micrographs and spectral analyses of arrays of nanostructures with varying diameters D and gaps g. Each 12 mm square in the array
consists of a square lattice of nanoposts of periodicity D þ g. a, Before metal deposition, the patterns of nanostructures show grey-scale variations but do not
display any colour. b, Following deposition of thin metal layers of uniform thickness, the full palette of colours is revealed. Nanostructure arrays with similar
colours are observed from bottom left to top right, indicating that areas with similar fill factors produce areas with similar colours. Structures with the
same periodicity display a wide range of colours (from bottom right to top left). The highlighted column was used to produce the image in Fig. 4b.
c, Selected experimental (i) and simulation (ii) spectra of nanodisks with g ¼ 120 nm. The trendlines approximate the movement of the peaks and dips with
varying sizes of the nanostructures. d, Correlation between dips and peaks observed in the experimental and simulation data.
sufficiently to be detected in an optical bright-field microscope,
regardless of their periodicity, as presented in the schematic.
Nanoposts with a height of 95 nm were patterned in a negativetone hydrogen silsesquioxane (HSQ) resist using electron-beam
lithography (EBL) as detailed in the Methods, after which a
chromium adhesion layer (1 nm), silver (15 nm) and a capping
layer of gold (5 nm) were deposited via electron-beam evaporation
(Fig. 1b, i–iii). We characterized the nanostructures before and
after metal deposition in a reflection bright-field microscope,
scanning electron microscope (SEM) and a microspectrophotometer. An SEM image of a small area of the nanostructures after
metal deposition is shown in Fig. 1b, iv.
2
To achieve a full palette of colours that span the visible range,
we systematically varied the diameters of the nanodisks (D) as well
as their interdisk separations ( g). The addition of metal layers of
uniform thickness transformed the grey-scale arrays of HSQ nanostructures in Fig. 2a into the brilliant display of colours in Fig. 2b
(viewed using reflection bright-field microscopy). From these arrays
of colours, three factors attest to the role of plasmon resonances
in colour formation: (i) colours were observed only upon the introduction of metal; (ii) equiperiodic regions (constant D þ g) traversing
the array diagonally from top left to bottom right did not exhibit
the same colour (unlike light diffraction off periodic structures);
and (iii) regions of similar fill factor (D/(D þ g)) have similar
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology
© 2012 Macmillan Publishers Limited. All rights reserved.
NATURE NANOTECHNOLOGY
a 5
(ii)
(i)
Reflectance (a.u.)
4
LETTERS
DOI: 10.1038/NNANO.2012.128
D = 90
(iii)
b
(i) 450 nm
(ii) 590 nm
(iii) 900 nm
5
D = 80
E/Einc
3
2
D = 70
D = 60
z
0
y
5
1
0
400
D = 50
S/Sinc
600
800 1,000
Wavelength (nm)
0
Figure 3 | Numerical simulation for structures of the same periodicity (D 1 g 5 120 nm). a, Simulated reflectance spectra plots for structures with varying
D. The corresponding experimentally observed colour is shown in the squares. Solid lines show reflectances for the combined structure (with disks and
backreflector), dotted lines for the case where metal nanodisks are removed, and dashed lines for the case where the backreflector is removed. Note that the
feature corresponding to a Fano resonance occurs at a constant wavelength of 900 nm for all values of D. Colour variation at constant periodicity can be
achieved only for the combined structure of nanodisks and backreflector. b, Electric field enhancement plots (top) and time-averaged power flow vector plots
(bottom) for a single repeat structure with D ¼ 90 nm at wavelengths of 450 nm (i), 590 nm (ii) and 900 nm (iii). Electric field enhancement calculated as
electric field (E) divided by the incident field (Einc), and the power flow vectors (S) were normalized by the incident Poynting vector (Sinc). Plane wave
illumination is incident from above in the z-direction and polarized along the y-axis. The spectral dip at 450 nm in a corresponds to light being absorbed by
both the metal structures and silicon substrate. The peak at 590 nm is due to plasmon resonance of the disk acting as a dipole antenna that re-radiates light
back to the observer. The inflexion point at 900 nm signifies a resonance where power flows around the disk, through the nanohole, and is absorbed by the
bottom rim of the nanohole array and substrate.
colours (most noticeably in the dark band going from bottom left to
mid right of the array), in accordance with the plasmon resonances
operating close to the quasi-static limit28, where retardation effects
are minimal and resonances are independent of size scaling.
Figure 2c, i presents spectral analyses of the rightmost column in
the array (indicated by a blue box), with the spectra exhibiting peaks
and dips that could be tuned across the visible spectrum by varying
D and thus the periodicity D þ g (see Supplementary Fig. S2 for
experimental spectra of other D, g values). Simulations (Fig. 2c, ii)
demonstrate qualitative agreement with the corresponding
experimental results, as is further exemplified in Fig. 2d, where
both peaks (triangles) and dips (squares) redshift with increasing
D (ref. 29). Through simulations, a subtle difference is found in
the origin of the spectral dips for D , 100 nm when compared
with larger disks (Supplementary Fig. S3). The dips for smaller
disks are due to power absorption by the disks and, to a lesser
extent, the backreflector. Together, the disk, post and backreflector
effectively act as an antireflection stack at this wavelength
(Supplementary Fig. S3c, iii and d, iii). Conversely, the dips for
larger disks are due to Fano resonances that result from the interference between the broad resonance of the nanoholes and nanodisks
with the sharp resonance of the surface modes19,20. At this resonance
condition, optical power flows around the nanodisks, through the
nanoholes, and is absorbed by the backreflector/silicon substrate
(see the Poynting vector plots in Supplementary Fig. S3d, ii). The
peaks correspond to the plasmon resonances of the disks, which
intensify for larger disks because of their increased scattering
strengths30 (Supplementary Fig. S4).
It is well known that the periodicity of nanoholes in a metal film
determines its optical resonance10,11, but it is less obvious how structures with a constant periodicity can produce a range of colours.
Figure 3 presents simulation results for structures with a constant
periodicity of 120 nm and D varying between 50 nm and 90 nm,
which exhibit multiple colours as indicated in the insets of
Fig. 3a. Simulations of the three different configurations shown in
Fig. 3a indicate that structures comprising only nanodisks or only
a backreflector plane cannot produce the colours observed. The metallic backreflector alone (dotted lines) displays a fairly constant spectrum across arrays with the same periodicity, with a point of inflexion
at 900 nm, the signature of a Fano resonance profile20, and a dip at
450 nm that is attributable to the antireflection stack at this wavelength, as described earlier. Evidence for the dip at 450 nm
having a different physical origin from the feature at 900 nm can be
seen in its invariance to changing periodicity (Supplementary
Fig. S5). For structures with just disks (dashed lines), a single peak
is observed corresponding to the nanodisk plasmon resonance that
blueshifts and intensifies with increasing D only within a narrow spectral range between l ¼ 570 nm and 590 nm. A span of colours is
achieved only in the combined structure; as the scattering strength
of the disks increases, the spectrum peak shifts in favour of the nanodisk resonance and away from the Fano resonance. Simulations
(Supplementary Fig. S6) indicate that structures consisting of disks
raised above a backreflector film without nanoholes would display
similar colours but without the Fano resonance. However, the effect
of Fano resonance in our system, although non-crucial, aids in narrowing the main spectral peaks for purer colours.
Figure 3b demonstrates the electric field enhancement and shows
the Poynting vector plots for the combined structure extracted at (i)
450 nm, (ii) 590 nm and (iii) 900 nm for D ¼ 90 nm. The nanodisk
appears to play different roles at each of these wavelengths: it is
absorbing in the antireflection stack in (i), scattering in (ii) and
enhancing absorption around the nanohole in (iii). These effects
can also be seen in the channelling of the Poynting vectors into
the nanodisk in (i), the strong fields signifying nanodisk plasmon
resonance in (ii), and the directing of power flow around the
nanodisk and into the base of the nanohole in (iii). The decreased
reflectance at l ¼ 900 nm compared to that of the backreflector
alone is evidence that the nanodisk acts as an antenna that enhances
the absorption around the nanohole23.
We now present the fabrication of microscopic colour images to
demonstrate the creation of arbitrary images with colour and tonal
control. The creation of a photo-realistic image is demonstrated in
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology
© 2012 Macmillan Publishers Limited. All rights reserved.
3
LETTERS
NATURE NANOTECHNOLOGY
a
b
d
DOI: 10.1038/NNANO.2012.128
c
e
Figure 4 | Full-colour image printing and resolution test patterns. a,b, Optical micrographs of the Lena image before (a) and after (b) metal deposition.
c, Optical micrograph of an enlarged region of the image, showing the remarkable detail and colour rendition on a micrometre-scale level. The accompanying
SEM image of the indicated region shows that the nanostructures that make up the image have a similar periodicity of 125 nm but exhibit different colours in
the optical image due to a small (30 nm) variation in nanodisk sizes. For clarity, the individual regions of similarly sized disks are separated by the dotted
lines. Each pixel consists of a 2 × 2 array of disks with a pitch of 250 nm. d,e, Resolution test patterns in the form of chequerboards approaching (d) and at
(e) the optical diffraction limit. In d, a set of chequerboards consists of a combination of two colours, one darker than the other. Each square of the
chequerboard is 375 nm in size and consists of an array of 3 × 3 structures, as shown in the corresponding SEM image. In e a similar chequerboard with
colour squares of 250 nm (with the size of pixels used in the Lena image) consists of an array of 2 × 2 structures, as shown in the SEM image.
The chequerboard pattern is only barely observable, even with a ×150 and 0.9 NA objective, demonstrating the patterning of colour pixels at the
optical diffraction limit. Scale bars: 10 mm (a, b), 1 mm (c), 500 nm (d, e).
the 50 × 50 mm square Lena image (previously presented as a
miniaturized grey-scale image31) shown in Fig. 4a–c. Colour information from bitmap images was coded pixel by pixel into the position, diameter (D) and separation ( g) of nanoposts formed in the
HSQ resist (for details see Supplementary Figs S7,S8). For this
demonstration, the pixel was a 250 × 250 nm square (that is, at
the theoretical resolution limit of the optical microscope, ×100
objective, numerical aperture (NA) ¼ 0.9, mid-spectrum wavelength of 500 nm; see http://www.microscopyu.com/articles/formulas/formulasresolution.html). Most pixels consisted of four
nanodisks, as shown schematically in Fig. 1b, ii, although single
nanodisks were also used to achieve the blue/purple colours.
The colour information latent in the grey-scale structures
(Fig. 4a) manifested upon deposition of the metal layers
(Fig. 4b). Remarkably, the resulting image closely reproduces the
details of the original image down to single-pixel elements, as
seen in the appearance of specular reflections in the eyes
(Fig. 4c). The accompanying SEM micrograph in Fig. 4c shows
the structures, which have the same centre-to-centre periodicity
but different D and g. This image is enlarged from the region indicated in Fig. 4c, which consists of four different colours.
To demonstrate the colour pixel resolution at the optical diffraction
limit, we patterned a set of chequerboard resolution test structures
with alternating colours. As shown in Fig. 4d,e, each square in the chequerboard consists of a 3 × 3 (Fig. 4d) or 2 × 2 (Fig. 4e) array of disks
per pixel. The centre-to-centre separation of the disks is 125 nm,
matching that of Fig. 4a–c. Although the number of disks per pixel
is reduced from nine to four disks in Fig. 4e, the colour scheme of
each chequerboard is preserved. The fact that the individual colours
are just resolved by a diffraction-limited optical microscope indicates
that the single pixels of four nanodisks were indeed able to support
individual colours at the optical diffraction limit.
The critical advantage of this approach is that the colour information latent in the resist structures can be replicated economically
4
onto multiple substrates using high-throughput methods (such as
NIL; Supplementary Fig. S9) once a master mould is fabricated.
In addition to achieving high resolution, the use of plasmonic resonators also provides secondary degrees of freedom to colour creation, including polarization dependence17. We anticipate that
further improvements in resolution and colour perception will be
achieved by using different geometries and/or smaller numbers of
nanostructures per pixel area.
In summary, we have presented an approach for full-colour
printing at the optical diffraction limit by encoding colour information into silver/gold nanodisks raised above a holey backreflector. The interplay of plasmon and Fano resonances, which can be
tuned by varying the size and separation of the nanodisks, results
in colours directly visible under a bright-field optical microscope.
These colours are preserved even when only four disks are present
in individual pixels of 250 × 250 nm squares, thus enabling
colour printing at a resolution of 100,000 d.p.i. This printing resolution brings us to the limit of visible-light imaging, where the
individual colour pixels are just barely resolvable using
diffraction-limited optics. Beyond obvious applications in high-resolution print image production, this method can also be used in
optical data storage and colour filters in lighting and
imaging technologies.
Methods
Electron-beam lithography. The figures in the main text show colour information
encoded into the dimension and position of nanostructures by EBL. The negativetone electron-beam resist HSQ (formulated as product number XR-1541-006, Dow
Corning) was spin-coated onto silicon substrates to a thickness of 95 nm. To avoid
thermally induced crosslinking of the resist, which would reduce its resolution, no
baking process was used. A computer-generated layout consisting of arrays of disks
with a range of nominal diameters (50–140 nm) and gaps (30–120 nm) was
designed. EBL was performed using an Elionix ELS-7000 EBL system with an
accelerating voltage of 100 kV and a beam current of 500 pA. The write field was set to
150 mm × 150 mm and the exposure step size was 2.5 nm. The dose used for the
nanodisk structures was 12 mC cm22. No proximity-effect correction was performed
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology
© 2012 Macmillan Publishers Limited. All rights reserved.
NATURE NANOTECHNOLOGY
DOI: 10.1038/NNANO.2012.128
for the exposure. Instead, to achieve well-defined nanodisk structures with steep
sidewalls, a high-contrast development process32 was used by developing samples in a
formulation of aqueous 1% NaOH, 4% NaCl in deionized (DI) water at 24 8C for
1 min, followed by rinsing under running DI water for 2 min, and isopropyl alcohol
(IPA). Finally, the samples were blow-dried under a steady stream of N2.
Nanoimprint lithography. Supplementary Fig. S8 shows the scaling up of colour
production over large areas using NIL as a means for creating the disk
nanostructures. The nanoimprint process was performed in an Obducat Sindre
600 thermal nanoimprinting system. A silicon mould (NIL Technology) with a
nanohole array (diameter, 100 nm; pitch, 200 nm; depth, 100 nm) occupying an area
of 1 cm × 1 cm was used as the mould to produce a large-area nanopost array. The
substrate for the nanoimprint was a 2 cm × 2 cm polycarbonate film with a thickness
of 125 mm. Before the imprinting process, the silicon mould was cleaned and a selfassembled monolayer of 1H,1H,2H,2H-perflourodecyltrichlorosilane (FDTS) antistiction coating was functionalized on the surface. This layer was necessary for
detachment of the mould from the imprinted substrate in the subsequent process.
The imprinting process was carried out at 150 8C under a pressure of 40 bar. This
condition was maintained for 300 s, after which the system was cooled to 30 8C
before manual detachment of the silicon mould from the imprinted
polycarbonate film.
Metal deposition. Metal deposition was performed using an electron-beam
evaporator (Explorer Coating System, Denton Vacuum). A chromium adhesion
layer (1 nm), a silver plasmon-active layer (15 nm) and a gold capping layer (5 nm)
were sequentially deposited onto the samples. All metals were deposited at a rate of
1Ås21. The chromium adhesion layer was necessary to provide scratch resistance in
the resulting metal layers and had minimal effect on the optical properties of the
structures. The gold capping layer hindered the sulphidation of silver. The working
pressure during evaporation was 1 × 1026 torr. The temperature of the sample
chamber was maintained at 20 8C during the entire evaporation process, with the
sample holder rotating at a rate of 50 r.p.m. to ensure uniformity of the deposition.
The fabricated structures were imaged using an Elionix ESM-9000 scanning
electron microscope with an accelerating voltage of 10 kV and a working distance
of 5 mm.
Optical measurements. To investigate the optical properties of the fabricated
structures, extinction spectra were measured in reflection mode using a QDI 2010
UV-visible-NIR range microspectrophotometer (CRAIC Technology). Both
incident and collected light were at normal incidence to the substrate, with the
electric field of the unpolarized light in plane with the substrate surface. Optical
micrographs were acquired using a Nikon MM-40/L3FA (Nikon) set-up with a
×100, 0.9 NA air objective as well as with a Olympus MX61 set-up with ×150,
0.9 NA air objective, using a JVC Colour Video Camera TKC14-81EG (JVC Corp)
for the former and an SC30 Olympus Digital Camera for the latter.
Numerical simulations. Simulations of the reflectance spectra were carried out
using a frequency domain solver from the Computer Simulation Technology
(CST AG) microwave studio commercial software. Unit-cell boundaries were used in
the plane of the disks and Floquet ports were used for terminating the domain in the
direction of incidence. The frequency-domain solver incorporated measured
wavelength dispersion of the permittivity of the various materials used in the
structure (data for silicon, gold and silver were taken from ref. 33 and HSQ was
simulated using a constant refractive index of 1.4).
Received 2 April 2012; accepted 29 June 2012;
published online 12 August 2012
LETTERS
10. Chen, Q. & Cumming, D. R. High transmission and low color cross-talk
plasmonic color filters using triangular-lattice hole arrays in aluminum films.
Opt. Express 18, 14056–14062 (2010).
11. Inoue, D. et al. Polarization independent visible color filter comprising an
aluminum film with surface-plasmon enhanced transmission through a
subwavelength array of holes. Appl. Phys. Lett. 98, 093113 (2011).
12. Xu, T., Wu, Y-K., Luo, X. & Guo, J. L. Plasmonic nanoresonators for highresolution colour filtering and spectral imaging. Nature Commun. 1, 1–5 (2010).
13. Abbe, E. A contribution to the theory of the microscope and the nature of
microscopic vision. Proc. Bristol Nat. Soc. 1, 200–261 (1874).
14. Huang, J., Wang, X. & Wang, Z. L. Controlled replication of butterfly wings for
achieving tunable photonic properties. Nano Lett. 6, 2325–2331 (2006).
15. Huo, F. et al. Polymer pen lithography. Science 321, 1658–1660 (2008).
16. Piner, R. D., Zhu, J., Xu, F., Hong, S. & Mirkin, C. A. ‘Dip-pen’ nanolithography.
Science 283, 661–663 (1999).
17. Ellenbogen, T., Seo, K. & Crozier, K. B. Chromatic plasmonic polarizers for
active visible color filtering and polarimetry. Nano Lett. 12, 1026–1031 (2012).
18. Laux, E., Genet, C., Skauli, T. & Ebbesen, T. W. Plasmonic photon sorters for
spectral and polarimetric imaging. Nature Photon. 2, 161–164 (2008).
19. Liu, H. & Lalanne, P. Microscopic theory of the extraordinary optical
transmission. Nature 452, 728–731 (2008).
20. Luk’yanchuk, B. et al. The Fano resonance in plasmonic nanostructures and
metamaterials. Nature Mater. 9, 707–715 (2010).
21. Ebbesen, T. W., Lezec, H. J., Ghaemi, H. F., Thio, T. & Wolff, P. A. Extraordinary
optical transmission through sub-wavelength hole arrays. Nature 391,
667–669 (1998).
22. Li, W-D., Ding, F., Hu, J. & Chou, S. Y. Three-dimensional cavity nanoantenna
coupled plasmonic nanodots for ultrahigh and uniform surface-enhanced
Raman scattering over large area. Opt. Express 19, 3925–3936 (2011).
23. Li, W-D., Hu, J. & Chou, S. Y. Extraordinary light transmission through opaque
thin metal film with subwavelength holes blocked by metal disks. Opt. Express
19, 21098–21108 (2011).
24. Liu, Y. J. et al. Optically tunable plasmonic color filters. Appl. Phys. A 107,
49–54 (2011).
25. Caldwell, J. D. et al. Plasmonic nanopillar arrays for large-area, high-enhancement
surface-enhanced Raman scattering sensors. ACS Nano 5, 4046–4055 (2011).
26. Chanda, D. et al. Coupling of plasmonic and optical cavity modes in quasi-threedimensional plasmonic crystals. Nature Commun. 2, 1–7 (2011).
27. Wells, S. M., Polemi, A., Lavrik, N. V., Shuford, K. L. & Sepaniak, M. J. Efficient
disc on pillar substrates for surface enhanced Raman spectroscopy. Chem.
Commun. 47, 3814–3816 (2011).
28. Maier, S. A. Plasmonics: Fundamentals and Applications (Springer, 2007).
29. Jain, P. K., Lee, K. S., El-Sayed, I. H. & El-Sayed, M. A. Calculated absorption and
scattering properties of gold nanoparticles of different size, shape, and
composition: applications in biological imaging and biomedicine. J. Phys. Chem.
B 110, 7238–7248 (2006).
30. Mie, G. Contributions to the optical characteristis of turbid tubes, particularly of
colloidal metal solutions (beitraege zur optik trueber Medien, speziell kolloidaler
Metalloesungen). Ann. Phys. 25, 377–445 (1908).
31. Yang, J. K. W., Duan, H., Law, J. B. K., Low, H. Y. & Cord, B. Miniaturization of
grayscale images. J. Vac. Sci. Technol. B 29, 06F313 (2011).
32. Yang, J. K. W. & Berggren, K. K. Using high-contrast salty development of
hydrogen silsesquioxane for sub-10-nm half-pitch lithography. J. Vac. Sci.
Technol., B 25, 2025–2029 (2007).
33. Palik, E. D. Handbook of Optical Constants of Solids Vol. 1, 386–350
(Elsevier, 1998).
Acknowledgements
References
1. Finlayson, C. E. et al. 3D bulk ordering in macroscopic solid opaline films by
edge-induced rotational shearing. Adv. Mater. 23, 1540–1544 (2011).
2. Haverinen, H. M., Myllylae, R. A. & Jabbour, G. E. Inkjet printing of light
emitting quantum dots. Appl. Phys. Lett. 94, 073108 (2009).
3. Kim, T-H. et al. Full-colour quantum dot displays fabricated by transfer
printing. Nature Photon. 5, 176–182 (2011).
4. Lee, S. Y. et al. Plasmon-enhanced structural coloration of metal films with
isotropic pinwheel nanoparticle arrays. Opt. Express 19, 23818–23830 (2011).
5. Ozaki, M., Kato, J. & Kawata, S. Surface-plasmon holography with white-light
illumination. Science 332, 218–220 (2011).
6. Xu, T., Wu, Y-K., Kaplan, A. F., Ok, J. G. & Guo, J. L. Structural colors: from
plasmonic to carbon nanostructures. Small 7, 3128–3136 (2011).
7. Ahn, S. H. & Guo, L. J. High-speed roll-to-roll nanoimprint lithography on
flexible plastic substrates. Adv. Mater. 20, 2044–2049 (2008).
8. Kustandi, T. S., Low, H. Y., Teng, J. H., Rodriguez, I. & Yin, R. Mimicking dominolike photonic nanostructures on butterfly wings. Small 5, 574–578 (2009).
9. Javidi, B. Optical and Digital Techniques for Information Security
(Springer, 2005).
This work was supported by the Agency for Science, Technology and Research (A*STAR)
Young Investigatorship (grant no. 0926030138) and SERC (grant no. 092154099).
The work made use of the SERC nano Fabrication, Processing and Characterization
(SnFPC) facilities in IMRE. The authors thank S.H. Goh, I.Y. Phang, J. Deng and V.S.F. Lim
for technical assistance, and M. Asbahi, M. Bosman, W.P. Goh and K.T.P. Lim (IMRE) and
K.K. Berggren (MIT) for fruitful discussions.
Author contributions
K.K., H.D. and J.K.W.Y. conceived the ideas and designed the experiments. K.K., H.D. and
J.K.W.Y. fabricated and characterized the samples. R.S.H., S.C.W.K. and J.N.W. performed
numerical simulations. All authors analysed the data and wrote the manuscript.
Additional information
Supplementary information is available in the online version of the paper. Reprints and
permission information is available online at http://www.nature.com/reprints. Correspondence
and requests for materials should be addressed to J.K.W.Y.
Competing financial interests
The authors declare no competing financial interests.
NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology
© 2012 Macmillan Publishers Limited. All rights reserved.
5