A first experimental demonstration of a planar superprism in silicon microphotonics technology using silicon on insulator (SOI) substrates is presented. Experimental results for anomalous wavelength-dependent angular dispersion in SOI triangular lattice planar photonic crystals are reported. An angular swing of 14° is measured for light propagating near the Γ-K direction as the input wavelength is changed from 1295 nm to 1330 nm, which corresponds to an angular dispersion of 0.4°/nm. For the Γ-M direction, a negative wavelength dispersion has been recorded. An opposite sign angular deviation of 21° is observed as the input wavelength is changed from 1316 nm to 1332 nm, i.e. a dispersion of 1.3°/nm.
©2004 Optical Society of America
The unusual dispersion properties of photonic crystals (PhC) outside their forbidden bandgaps are attracting increasing attention [1–4]. One particular example is the “superprism” effect , related to group velocity dispersion. A large change in the deflection angle of a light beam within the photonic crystal is achieved by a slight change of the wavelength or of the incident angle. This “superprism” effect has been recently demonstrated by Kosaka et al. [3,4] in their “autocloned” three-dimensional PhCs as well as by Wu et al [6,7] and Baumberg  in planar photonic microstructures in III–V semiconductors.
The aim of the present work is to investigate the “superprism” effect in a planar 2D geometry compatible for integration with SOI compact lightwave circuits. The angular dispersion is studied for several incident angles of a guided beam. This corresponds to different directions of the Brillouin zone in a slab two dimensional PhC structure which is realized in SOI technology for the first time.
2 Approach developed
Our system consists of a SOI film perforated by a triangular PhC lattice of circular air holes. The top silicon layer of the SOI wafer is 240 nm thick and the buried oxide thickness is 1 µm. The PhC lattice parameter and hole diameter are respectively, 460 nm and 170 nm, i.e. a filling ratio of 12%. Top cladding of the structure is air. The devices are fabricated in a standard silicon microfabrication process line, compatible with very large-scale integrated electronics, on SOI wafers. The samples were fabricated using 193 nm deep-UV lithography and reactive ion etching process (RIE). This process allowed fabricating high quality samples with very low side wall roughness. This is illustrated in figure 1 which is a SEM view of a SOI waveguide fabricated at the same time and used for the input and output guided beams.
The experimental device realized to test the “superprism” effect is shown in Fig. 2. The PhC area fills a 40 µm radius semicircle inside a 100 µm diameter slab region. In order to determine the PhC dispersion properties near different points of the Brillouin zone, the device includes several SOI input waveguides with incident angles ranging from -60° to +60° every 15°. The 10 µm input waveguide width provides a reasonably small beam divergence (≈3°) for light propagating through the 100 µm slab region. Special care was taken in the design to insure light propagation only in the fundamental mode in the input waveguides. This point was especially critical since at a 10 µm width the waveguides are intrinsically multimode. The light transmitted through the slab area is collected by a set of 3 µm wide output waveguides that are spaced 3.5° apart.
3. Experimental results
The experimental characterization is based on a set-up using a tunable semiconductor laser in the spectral range between 1250 and 1350 nm. A linearly polarized light beam is coupled into an input waveguide using a polarization maintaining lensed-fiber. The output light is collected by a 40× objective and is either imaged with an IR vidicon camera or measured with a large area IR detector. An iris aperture placed before the IR detector is used to spatially filter out any scattered light and guarantee that only the light from the waveguide output is reaching the photodetector.
The experiments performed for the different incident angles allowed to observe wavelength dependent angular dispersion phenomena at 30° and 45° input angles for TM light polarization. IR vidicon camera photographs of the output facet at different wavelength for these incident angles are displayed in Fig. 3(a) and 3(b), respectively. Each spot corresponds to one of the output waveguides that are spaced 3.5° apart. Since the width of the input waveguides is 10 µm while output waveguides are 3 µm wide, most of the light intensity is usually concentrated in two or three output waveguides. The strong intensity contrast between lightened and “dark” waveguides demonstrates the absence of significant coupling between adjacent output waveguides in the star coupler zone (Fig. 3(a)).
The observed images clearly show a variation of the light deflection with the wavelength. This variation is however quite different for 30° and 45° incidence angles. For the 30° input angle, the range of output angles is also near 30°. This angle corresponds to the Γ-K direction in a triangular lattice PhC. The deflection angle increases with wavelength. This situation is similar to the one usually encountered when light penetrates from a more dense into a less dense optical media with a positive wavelength dispersion of the refractive index. Note that the effective index in the PhC area is normally lower than in the rest of the slab region.
For the 45° input angle, the output angle range is around 56°. This is not far from the Γ-M direction (60°) in a triangular lattice PhC. In this case, the refraction angle decreases with wavelength, which corresponds to a negative wavelength dispersion of the refraction index. The variation of the angular deflection with wavelength is higher for the 45° input angle than for the 30° one, but transmission losses are also higher.
These qualitative evaluations of the angular dispersion and propagation losses are confirmed by quantitative measurements. The overall transmission losses at 30° and 45° input angles have been determined. In these measurements, light from several outputs is integrated. The losses are normalized with respect to the transmission of a blank superprism structure, in which the silicon slab between the input and output waveguides does not include any etched PhC structure. For the Γ-K direction, at the wavelength corresponding to the maximum of the superprism transmission, which is around 1.325 µm, the losses related to the propagation through the 40 µm long PhC are approximately 4 dB. They are considerably lower than those reported in other superprism experiments [6,7].
For the 45° input angle, propagation losses are roughly one order of magnitude higher, around 12 dB. This result is not surprising since when looking to the PhC structure (Fig. 2(b)), stronger reflection and scattering is expected for the light traveling in the Γ-M direction.
To display good performances, a wavelength selective filter should exhibit simultaneously both high wavelength dispersion and narrow bandwidth. The evaluation of these parameters from our experiments is however somewhat complicated by the presence of marked oscillations in the transmission spectra. These oscillations are possibly related to interference effects due to periodic non-uniformities in the PhC region .
In order to determine more accurately the wavelength position of the transmission maximum from the spectral dependences, they are approximated by a gaussian profile. Such a profile becomes a simple parabola function when represented on a logarithmic scale. This gaussian approximation allows also to evaluate the spectral bandwidth, defined at -10 dB level from transmission maximum of each output spectrum.
The transmission spectra show a well pronounced angular dispersion effect, especially at 45° input angle. The wavelength positions of the spectral maximum as a function of the output deflection angle are plotted in Figs. 6(a) and 6(b) for the 30° and 45° input angles, respectively. An angular swing of 14° is observed for the light propagating near the Γ-K direction as the input wavelength is changed from 1.295 µm to 1.33 µm, which gives an angular dispersion of 0.4°/nm. For the Γ-M direction, an opposite sign deviation of 21° is observed as the input wavelength is changed from 1.316 µm to 1.332 µm, meaning an angular dispersion of 1.3°/nm. Such a strong angular dispersion is two orders of magnitude higher than for a conventional prism.
Relatively large spectral bandwidth values are observed in our experiments. Near Γ-K directions, the bandwidth is in an interval between 40 nm and 58 nm. Near Γ-M directions, this interval is substantially the same: 30 nm–58 nm, but it is curious to mention that that bandwidth is smaller in the middle of the angular span. For the Γ-K direction the bandwidth increases with the wavelength.
Furthermore, the transmission level is higher in the middle of the angular span and approximately one order of magnitude lower at the extremities of the span. This situation is common both for Γ-K and Γ-M directions. The reasons of such behavior are not clear at the moment. In fact the observed transmission spectra (Figs. 4 and 5), represents a convolution of two distinct wavelength-dependent phenomena. One is related to the variation of the deflection angle with the wavelength and another one with the transmission variation with the wavelength for some fixed particular angles. While the angular dispersion can be obtained from the iso-frequency contours of a PhC in a plane wave approximation, the transmission through a finite size PhC requires additional modeling work that has not been performed at the moment.
To summarize, anomalous wavelength-dependent angular dispersion in a SOI triangular lattice planar PhC was studied for different directions of the Brillouin zone. A strong angular dispersion of 0.4°/nm was obtained for the Γ-K direction and a negative wavelength dispersion up to 1.3°/nm was found for the Γ-M direction. When compared with previous works, the measured dispersion is lower than the one obtained with 3D PhC [1–3], which was about 3°/nm, but it is larger than the one obtained with slab PhC in III–V semiconductors (0.5°/nm) [6, 7]. To our knowledge, this is the first experimental demonstration of a planar superprism in the silicon photonic technology using silicon on insulator (SOI) substrates.
The authors are very grateful to Dr. Abderrahim Ramdane for the critical reading of the manuscript and the useful comments. This work was supported by the ACI Lambdaconnect project.
2. M. Loncar, D. Nedeljkovic, T. Doll, J. Vuckovic, A. Scherer, and T. Pearsall, “Waveguiding in planar photonic crystals,” Appl. Phys. Lett. , 77, 1937–1939, (2000). [CrossRef]
3. H. Kosaka, T. Kawashima, A. Tomita, M. Notomi, T. Tamamura, T. Sato, and S. Kawakami, “Superprism phenomena in photonic crystals,” Phys. Rev. B , 58, 10096–10099, (1998). [CrossRef]
4. H. Kosaka, T. Kawashima, A. Tomita, M. Notomi, T. Tamamura, T. Sato, and S. Kawakami, “Photonic crystals for micro lightwave circuits using wavelength- dependent angular beam steering,” Appl. Phys. Lett. , 74, 1370–1372, (1999). [CrossRef]
5. T. Baba and M. Nakamura, “Photonic Crystal Light Deflection Devices Using the Superprism Effect,” IEEE Journ. of Quant. Electron. , 38, 909–914, (2002). [CrossRef]
6. L. Wu, M. Mazilu, T. Karle, and T. F. Krauss, “Superprism Phenomena in Planar Photonic Crystals,” IEEE Journ. of Quant. Electron. , 38, 915–918 (2002). [CrossRef]
7. L. Wu, M. Mazilu, J.-F. Gallet, and T. F. Krauss, “Square lattice photonic crystal collimator,” Photonic and Nanostructures , 1, 31–36 (2003). [CrossRef]
8. J. J. Baumberg, N. M. B. Perney, M. C. Netti, M. D. C. Charlton, M. Zoorob, and G. J. Parker, “Visible-wavelength super-refraction in photonic crystal superprisms,” Appl. Phys. Lett. , 85, 354–356 (2004). [CrossRef]
9. C. J. M. Smith, R. M. De La Rue, M. Rattier, S. Olivier, H. Benisty, C. Weisbuch, T. F. Krauss, R. Houdré, and U. Oesterle, “Coupled guide and cavity in a two-dimensional photonic crystal,” App. Phys. Lett. 78, 1487–1489 (2001). [CrossRef]