Вы находитесь на странице: 1из 8

Metal-assisted photonic mode for ultrasmall

bending with long propagation length at visible


wavelengths
Chengyuan Yang,1 Ee Jin Teo,2* Tian Goh,1 Siew Lang Teo,2 Jing Hua Teng,2 and
Andrew A. Bettiol1,4
1

Center for Ion Beam Applications, 2 Science Drive 3, Department of Physics, National University of Singapore,
117542 Singapore
2
Institute of Materials Research and Engineering, A*STAR (Agency for Science, Technology and Research), 3
Research Link, 117602 Singapore
4
Co-corresponding author: phybaa@nus.edu.sg
*
teoej@imre.a-star.edu.sg

Abstract: In this work, we investigate the use of metal-assisted photonic


guiding in a polymer-metal waveguide as an alternative approach for high
density photonic integration at visible wavelengths. We demonstrate high
confinement and long propagation length in sub-wavelength dimensions
down to 300nm 200nm using leakage radiation microscopy at a
wavelength of 632.8 nm. Simulations using the finite element method
(FEM) show that the optimum dimension that gives good confinement and
propagation length is similar to that of the predicted plasmonic mode
supported in the same waveguide. Under such optimum conditions, the
metal-assisted photonic mode shows a five times longer propagation length
and higher transmission efficiency for all 90 bending radius down to 1 m
compared to the plasmonic mode.
2012 Optical Society of America
OCIS codes: (240.6680) Surface plasmons; (250.5300) Photonic integrated circuits; (250.5460)
Polymer waveguides.

References and links


1.
2.

G. T. Reed and A. P. Knights, in Silicon Photonics: An Introduction (Wiley, England, 2004).


R. Zia, A. J. Schuller, and M. L. Brongersma, Near-field characterization of guided polariton propagation and
cutoff in surface plasmon waveguides, Phys. Rev. B 74(16), 165415 (2006).
3. E. Ozbay, Plasmonics: Merging photonics and electronics at nanoscale dimensions, Science 311(5758), 189
193 (2006).
4. J. C. Weeber, A. Dereux, C. Girard, J. R. Krenn, and J. P. Goudonnet, Plasmon polaritons of metallic nanowires
for controlling submicron propagation of light, Phys. Rev. B 60(12), 90619068 (1999).
5. S. I. Bozhevolnyi, V. S. Volkov, E. Devaux, and T. W. Ebbesen, Channel plasmon-polariton guiding by
subwavelength metal grooves, Phys. Rev. Lett. 95(4), 046802 (2005).
6. D. F. P. Pile, T. Ogawa, D. K. Gramotnev, Y. Matsuzaki, K. C. Vernon, K. Yamaguchi, T. Okamoto, M.
Haraguchi, and M. Fukui, Two-dimensionally localized modes of a nanoscale gap plasmon waveguide, Appl.
Phys. Lett. 87(26), 261114 (2005).
7. L. Liu, Z. Han, and S. He, Novel surface plasmon waveguide for high integration, Opt. Express 13(17), 6645
6650 (2005).
8. S. A. Maier, P. G. Kik, H. A. Atwater, S. Meltzer, E. Harel, B. E. Koel, and A. A. G. Requicha, Local detection
of electromagnetic energy transport below the diffraction limit in metal nanoparticle Plasmon waveguides, Nat.
Mater. 2(4), 229232 (2003).
9. J. Grandidier, S. Massenot, G. C. desFrancs, A. Bouhelier, J.-C. Weeber, L. Markey, A. Dereux, J. Renger, M. U.
Gonzalez, and R. Quidant, Dielectric-loaded surface plasmon polariton waveguides: Figures of merit and mode
characterization by image and fourier plane leakage microscopy, Phys. Rev. B 78(24), 245419 (2008).
10. B. Steinberger, A. Hohenau, H. Ditlbacher, F. R. Aussenegg, A. Leitner, and J. R. Krenn, Dielectric stripes on
gold as surface plasmon waveguides: Bends and directional couplers, Appl. Phys. Lett. 91(8), 081111 (2007).
11. T. Holmgaard, S. I. Bozhevolnyi, L. Markey, and A. Dereux, Dielectric-loaded surface plasmon-polariton
waveguides at telecommunication wavelengths: Excitation and characterization, Appl. Phys. Lett. 92(1),
011124 (2008).

#173304 - $15.00 USD

(C) 2012 OSA

Received 26 Jul 2012; revised 17 Sep 2012; accepted 27 Sep 2012; published 3 Oct 2012

8 October 2012 / Vol. 20, No. 21 / OPTICS EXPRESS 23898

12. T. Holmgaard, Z. Chen, S. I. Bozhevolnyi, L. Markey, A. Dereux, A. V. Krasavin, and A. V. Zayats, Bend- and
splitting loss of dielectric-loaded surface Plasmon-polariton waveguides, Opt. Express 16(18), 1358513592
(2008).
13. A. V. Kravasin and A. V. Zayats, Passive photonic elements based on dielectric-loaded surface plasmon
polariton waveguides, Appl. Phys. Lett. 90(21), 211101 (2007).
14. V. J. Sorger, Z. Ye, R. F. Oulton, Y. Wang, G. Bartal, X. Yin, and X. Zhang, Experimental demonstration of
low-loss optical waveguiding at deep sub-wavelength scales, Nature Comm. 2, 331 (2011).
15. H.-S. Chu, E.-P. Li, P. Bai, and R. Hedge, Optical performance of single-mode hybrid dielectric-loaded
plasmonic waveguide-based components, Appl. Phys. Lett. 96(22), 221103 (2010).
16. W. L. Barnes, A. Dereux, and T. W. Ebbesen, Surface plasmon subwavelength optics, Nature 424(6950), 824
830 (2003).
17. P. Berini, Plasmon-polariton waves guided by thin lossy metal films of finite width: Bound modes of symmetric
structures, Phys. Rev. B 61(15), 1048410503 (2000).
18. D. Dai, B. Yang, L. Yang, Z. Sheng, and S. He, Compact microracetrack resonator devices based on small SU-8
polymer strip waveguides, IEEE Photon. Technol. Lett. 21(4), 254256 (2009).
19. S. Massenot, J. Grandidier, A. Bouhelier, G. Colas des Francs, L. Markey, J.-C. Weeber, A. Dereux, J. Renger,
M. U. Gonzlez, and R. Quidant, Polymer-metal waveguides characterization by fourier plane leakage radiation
microscopy, Appl. Phys. Lett. 91(24), 243102 (2007).
20. www.comsol.com
21. E. D. Palik, Handbook of Optical Constants and Solids (Academic, 1985).
22. W. Wang, Q. Yang, F. Fan, H. Xu, and Z. L. Wang, Light propagation in curved silver nanowire plasmonic
waveguides, Nano Lett. 11(4), 16031608 (2011).

1. Introduction
Photonic integrated circuits (PIC) can potentially play an important role in the next generation
of devices for telecommunications, information processing and biological and chemical
sensing applications. In order to achieve high packing density on a single chip, it is important
to be able to route signals around tight bends. One approach is to use a high-index-contrast
system such as silicon-on-insulator to obtain high confinement [1]. Recently, there has been
an intense effort in using plasmonics for tight confinement and localization of light below its
diffraction limit. Basically, surface plasmon polaritons (SPP) are light waves coupled to
electron oscillations that can propagate along a metal/dielectric interface. Several types of
SPP waveguides have been proposed including metallic wires [24], grooves [5], metalinsulator-metal [6,7], and metallic nanoparticle waveguides [8]. Dielectric loaded plasmonic
waveguides have also been experimentally and theoretically studied in great detail and their
potential for integrated photonics has been shown from the bending loss and coupling length
[913]. More recently, hybrid plasmonic waveguides have received more attention due to
their superior sub-wavelength confinement [14,15].
Most studies on SPP waveguides have concentrated on using near infrared light for
excitation, with the aim of combining the high bandwidth of photonics with the small
footprint of electronics for compact and fast processors on a silicon platform. Although ultrahigh confinement can be obtained with plasmonic waveguides, propagation length is very low
due to high ohmic loss in metals. This becomes worse with decreasing wavelength [16] as the
fraction of modal power in the metal increases [17]. This limits its applications in nanofluidics
devices and biosensors which require visible wavelengths for excitation and detection of
fluorescent markers.
In this paper, we investigate the use of metal-assisted photonic guiding near cut-off limit
in a polymer-metal waveguide. This design relies on the high confinement of metal and the
low propagation loss of dielectric for long sub-wavelength propagation at visible range.
Polymer is a good choice of material due to its high transparency, low cost, ease of fabrication
and scalability. It is also a good host medium that can be easily doped with nonlinear
materials for gain or active waveguides. However, the bending radius of conventional
polymer waveguides is in the range of hundreds of microns due to significant light leakage
into the quartz substrate [18]. We are able to achieve high index contrast by using the metal as
a low index substrate to increase light confinement to a region beyond the cut-off limit of a
polymer-on-quartz waveguide. Silver is chosen as the metallic layer due to its lower damping

#173304 - $15.00 USD

(C) 2012 OSA

Received 26 Jul 2012; revised 17 Sep 2012; accepted 27 Sep 2012; published 3 Oct 2012

8 October 2012 / Vol. 20, No. 21 / OPTICS EXPRESS 23899

loss at visible wavelengths. In the same structure, plasmonic guiding at the silver/polymer
interface can be excited simply by changing the polarization of the incident light. Previously,
this mode has been studied extensively at near infrared region [913]. Here, we have
compared the optical performance of both the plasmonic and photonic modes in terms of their
propagation lengths, effective index, minimum supported size and 90 bending loss at 632.8
nm wavelength.
2. Effective index, propagation length and confinement factor
A negative resist, maN, with a refractive index of 1.64 is deposited onto a 100 nm thick silver
surface that has been electron beam evaporated onto a quartz substrate. Electron beam
lithography is then used to pattern waveguide structures onto the resist. Characterization of
these nanophotonic waveguides is carried out using leakage radiation microscopy (LRM). A
wavelength of 632.8 nm from a HeNe laser is used to both excite the supported modes and
couple them into waveguides at normal incidence. The laser light is focused onto the sample
using a 40 objective (Olympus UPlanFLN 40 , NA 0.75) and scattered off the edge of the
waveguide taper. This generates a continuum of wavevectors, some of which will satisfy the
phase matching condition, resulting in a guided mode [19]. A polarizing beam splitter and a
half-wave plate are inserted into the incoming beam path, enabling discrimination between the
TE and TM polarizations. In this case, the polarization aligned parallel and perpendicular to
the substrate are the TE mode and TM modes respectively. By using index matching fluid
through a high-numerical aperture lens of 1.4 (Zeiss Plan-Apochromat 63 ), we are able to
collect the light leakage from the waveguide. Our microscope permits access to both the direct
and Fourier plane. The direct plane image collected from the camera is used to determine the
propagation length of the waveguide, while the Fourier plane image is used to determine the
effective index of the guided modes. Simulations are carried out using COMSOL 4.2 Finite
Element Method (FEM) software [20].
Figure 1(a) shows a schematic diagram of the polymer-metal waveguide configuration.
We have considered a dielectric load thickness of 200 nm ~ / (2nw ) . At this thickness, a
conventional polymer waveguide without the metal layer will have no guided photonic
modes. With the addition of 100 nm of silver, negligible light is leaked into the substrate. The
high reflectivity and very low refractive index of silver (nAg = 0.1346 + 3.9882i) at 632.8 nm
wavelength causes an abrupt change of the refractive index and the optical mode profile at the
polymer/metal interface [21]. This greatly increases the confinement in the vertical direction
and a fundamental photonic mode can be seen in the TE polarization (TE0) with an effective
index of 1.15 for a width of 500 nm [Fig. 1(b)]. A plasmonic mode with higher effective
index of 1.69 can be obtained in TM polarization (TM0) [Fig. 1(c)].

Fig. 1. (a) Schematic cross-sectional view of the waveguide. (b) Electric field distribution of
the fundamental (b) photonic TE0 and (c) plasmonic TM0 modes supported in the waveguide.
The waveguide size is 500 200 nm in the simulation.

#173304 - $15.00 USD

(C) 2012 OSA

Received 26 Jul 2012; revised 17 Sep 2012; accepted 27 Sep 2012; published 3 Oct 2012

8 October 2012 / Vol. 20, No. 21 / OPTICS EXPRESS 23900

The waveguide modes are excited from the edge of the taper as shown in Fig. 2(a). The
direct plane images in Fig. 2(b) show modes propagating from left to right in the waveguides
for widths of 300, 400 and 500 nm. The intensity distribution in the waveguides shows that
the waveguides are single mode. Multimode propagation is evident for widths greater than
600 nm, with the appearance of a beating pattern in the waveguides and an additional line of
lower index in the Fourier image. As the width is reduced to 300 nm, significant leakage of
light can be seen along the waveguide as the mode becomes weakly guiding. We are able to
determine the effective index of the mode after calibration of the large central band in Fourier
space with the numerical aperture of the excitation objective [Fig. 2(c)]. The guided TE0
mode can be seen as a bright vertical line, with an effective index 1.15 for waveguide of 500
200 nm. A bright ring with an index of 1.07 can also be seen in the image due to the
excitation of the surface plasmon at the Ag/air interface outside the waveguide.

Fig. 2. (a) Scanning electron micrograph (SEM) showing the excitation of the modes at the
edge of the waveguide taper. Light propagation in the waveguide (b) in direct plane for widths
of 300, 400 and 500 nm. (c) Fourier image recorded by LRM of the guided photonic mode for
a 500 200 nm waveguide.

Figures 3(a) and 3(b) show plots of the effective index and propagation length as a
function of width for the fundamental TM0 (Plasmonic) and TE0 (Photonic) modes. The
effective index of TE0 mode shows good agreement between the experimental data measured
using the Fourier image and simulations performed using FEM (nAg = 0.1346 + 3.9882i) [21].
Much higher mode index is calculated for the TM0 modes, which cannot be detected with our
collection objective for widths greater than 160 nm. Hence, we have only studied the
plasmonic modes theoretically. From the direct images, we are able to determine the
propagation length of the TE0 mode, defined as the length at which the intensity decreases to
1/e of its initial value, by fitting the intensity decay with an exponential curve. Guided modes
can be seen for widths down to 300 nm. The discrepancy between experimental and
calculated data for widths larger than 600 nm is most likely due to the existence of higher
order modes at these widths. For the calculated TM0 mode, the propagation length is low and
rather constant at about 3.5 m, before increasing at widths smaller than 100 nm. A
comparison between the modes shows that TE0 mode has a much lower propagation loss than
the TM0 mode. For a width of 500 nm, the propagation loss of the TE0 photonic mode (~0.25
dB/m) is about five times smaller than the TM0 mode (~1.25 dB/m). The major factors
contributing to the loss are ohmic losses and scattering due to surface roughness. The TM0
mode has a higher ohmic loss because most of the energy resides in the metal rather than the
polymer structure.

#173304 - $15.00 USD

(C) 2012 OSA

Received 26 Jul 2012; revised 17 Sep 2012; accepted 27 Sep 2012; published 3 Oct 2012

8 October 2012 / Vol. 20, No. 21 / OPTICS EXPRESS 23901

Fig. 3. Plots of the (a) mode index, (b) propagation length and (c) confinement factor as a
function of width. Solid red and black lines correspond to simulated data for plasmonic TM0
mode and photonic TE0 mode respectively. The black square symbols correspond to
experimental data for TE0 mode. The broken vertical line indicates the cut-off width for the
TE0 mode.

When designing photonic integrated circuits, it is important to choose optimum


parameters that give maximum transmission of signals over sharp bends. Figure 3(c) shows
the confinement factors simulated at different widths for the photonic and plasmonic modes.
Confinement factor is defined as the fraction of the total power that is confined in the
waveguide structure. A maximum confinement of 85% is reached for TM0 mode at a width of
300 nm. As the width drops below 200 nm, the confinement decreases sharply and the mode
extends substantially out of the dielectric structure. This causes the propagation length to
increase as the interaction with metal decreases with poorer confinement. Therefore the tradeoff between optical confinement and metallic losses prohibits long propagation with ultrahigh confinement with plasmonic guiding. For the TE0 modes, we see lower confinement
values that saturate at around 65% at 500 nm. Here, photonic guiding for widths from 300 nm
to 600 nm shows that the propagation length increases with increasing confinement,
indicating that it is possible to obtain sufficiently high confinement and low propagation loss
at the same time. As the width increases further, the portion of the mode that interacts with the
metal increases, thus decreasing the propagation length. In the next section, we have

#173304 - $15.00 USD

(C) 2012 OSA

Received 26 Jul 2012; revised 17 Sep 2012; accepted 27 Sep 2012; published 3 Oct 2012

8 October 2012 / Vol. 20, No. 21 / OPTICS EXPRESS 23902

compared the transmission efficiency and bending loss of both photonic and plasmonic
guiding at the optimum width that gives good confinement and propagation length. This is
500 nm for the photonic mode and 300 nm for the plasmonic mode.
3. Transmission and bending loss
We have experimentally studied the transmission and bending loss of the photonic mode with
a radius of curvature from 5 m down to 0.5 m. Shown in Fig. 4(a) are the Atomic Force
Microscopy (AFM) images of a 90 and U-bend with 1 m radius of curvature. In Fig. 4(b),
we show the LRM images of light propagating around the 90 and U-bends for 1, 3 and 5 m
radii of curvature. It can be seen that there is significant transmission even at a radius of 1 m.

Fig. 4. (a) AFM of 90 and U-bend with 1 m radius of curvature. (b) LRM of TE0
transmission around 90 and U-bends with 1, 3 and 5 m radii of curvature.

The transmission curve for the 90 bend has been extracted from the input and output of
the bend as shown in Fig. 5(a). The bending loss can be extracted from the transmission curve
after subtracting the propagation loss around the bend (dB/bend) [Eq. (1)].

dB r
dB
Bending loss
.
= 10 log T propagation loss
Bend
m 2

(1)

Here, T is the transmission around the bend and r is the radius of the bend. The propagation
loss (dB/m) is assumed to be the same in the bend and straight sections of the waveguide. As
the radius of curvature reduces, the TE0 transmission initially increases until it reaches a
maximum of 70% at 3 m radius of curvature, before reducing again [Fig. 5(b)]. For radius
>3 m, the bending loss is rather constant and less than 0.5 dB/bend [Fig. 5(c)]. In this
regime, the linear increase in propagation loss per bend dominates over the slight decrease in
bending loss, resulting in an overall reduction in transmission with increasing radius. For
radius <3 m, the rapid increase in bending loss dominates the linear decrease in propagation
loss per bend, causing a sharper reduction in transmission with smaller radius. According to
the 3D FEM simulations, the TE0 transmission is in good agreement with experimental data.
Also shown in Figs. 5(b) and 5(c) are the simulated transmission and bending losses,
respectively for the TM0 mode for a width of 300 nm. As the radius reduces, the transmission
increases until it reaches a maximum of 50% at 1 m before reducing again. Here, the TM0
bending loss is lower than that of TE0 mode. However, the much higher propagation loss per
#173304 - $15.00 USD

(C) 2012 OSA

Received 26 Jul 2012; revised 17 Sep 2012; accepted 27 Sep 2012; published 3 Oct 2012

8 October 2012 / Vol. 20, No. 21 / OPTICS EXPRESS 23903

bend dominates over the bending loss, resulting in lower transmission for all radii of curvature
down to about 1 m. Only when the bending radius becomes much shorter than the
propagation length can we observe higher transmission for the TM0 mode. Direct comparison
of the TE0 bending losses with other types of SPP waveguides is difficult due to the lack of
published data measured at 632.8 nm. One example is Ag nanowires formed by chemical
synthesis [22] where the bending loss only reduces to 1dB at radius of curvature greater than
30 m, compared to 2 m in our case. Similar transmission data are predicted for hybrid
dielectric loaded plasmonic waveguide at 1550 nm [15].

Fig. 5. (a) Close-up of the experimental and calculated transmission around a 1 m bending
radius. Plot of the transmission (b) and bending loss (c) data for the observed TE0 and the
predicted TM0 modes.

4. Conclusion

Using a 100 nm thick Ag layer as a lower cladding layer, we have demonstrated subwavelength visible light guiding down to 300 200 nm with ultra-small bending radius in a
polymer waveguide. It is found that the optimum width that gives good confinement and
propagation length is 500 nm for photonic mode and 300 nm for plasmonic mode. Under such
conditions, a propagation loss of 0.25 dB/m and 1.25 dB/m is observed for photonic and
predicted for the plasmonic mode respectively. A total transmission efficiency of 70% (1.8
dB) with a bending loss of 0.5 dB/bend can be obtained for TE0 mode at a 90 bending radius
of 3 m. Even though the plasmonic mode shows lower bending loss, the photonic mode still
has higher transmission for all radii of curvature down to 1.0 m. This means that by having a
slightly bigger polymer-metal waveguide, we are able to support the photonic mode in
addition to the plasmonic mode, which has a longer propagation length and higher
transmission around sharp bends. Both of these factors are important for high density
photonics circuitry. By replacing the polymer in this system with other high refractive index
semiconductor or dielectric materials, we can obtain even higher confinement or other
functionalities.
Besides passive waveguides, this metal-assisted photonic mode holds great promise for
active devices in signal modulation, nanolasers and amplifiers. Since the mode is highly
confined to the polymer rather than at the interface, the overlap of the photonic mode with the
active medium is much greater than that which can be achieved with plasmonic guiding. This
#173304 - $15.00 USD

(C) 2012 OSA

Received 26 Jul 2012; revised 17 Sep 2012; accepted 27 Sep 2012; published 3 Oct 2012

8 October 2012 / Vol. 20, No. 21 / OPTICS EXPRESS 23904

means that it is much more sensitive to any change in dielectric properties with external
control. According to FEM simulations, a gain coefficient of 400 cm1 is able to achieve
lossless propagation for the photonic mode compared to a much higher threshold of 2200
cm1 for the plasmonic mode. This can be easily achieved by doping the polymer with laser
dyes such as Rhodamine B. By making use of the best aspects of plasmonics and photonics,
such system can potentially offer a platform for nanoscale photonic integrated circuits.
Acknowledgments

This work was supported by the A*STAR Metamaterials-Nanoplasmonics Research Program


under Grant No. A*STAR-SERC 0921540098 and the Ministry of Education (Singapore)
under the Academic Research Fund Tier 1 grant (R-144-000-291-112).

#173304 - $15.00 USD

(C) 2012 OSA

Received 26 Jul 2012; revised 17 Sep 2012; accepted 27 Sep 2012; published 3 Oct 2012

8 October 2012 / Vol. 20, No. 21 / OPTICS EXPRESS 23905

Вам также может понравиться