Research Review ’03-’05 (2006年4月18日刊行)

Preface

This is the third research review of Baba Laboratory, which summarizes activities and newly equipped facilities from 2003 to 2005. In previous two issues, papers were roughly categorized into those for photonic crystals (PCs), microdisks and silicon photonics. In this issue, small items such as nanolaser, slow light, and negative refractive optics are added, as they have become novel keywords for PC physics and applications.

Regarding PCs, two big events PECS V and VI (Int. Sympo. Photonic and Electromagnetic Crystal Structures) were held at Kyoto and Crete, respectively. In the preceding four symposia, main topics were the PC physics such as the photonic bandgap and basic technologies such as the PC slab. On the other hand, the recent two focused on how to use PCs and how effective they are in other physics areas and in practical applications. Such a new trend was also summarized in Roadmap on Photonic Crystals edited by Prof. S. Noda and myself and published from Kluwer Academic in 2003.

PC nanolasers and nanocavities have attracted attention of worldwide researchers because of their record high Q factor and small modal volume V. In 2002, we obtained the photopumped lasing in GaInAsP PC lasers in a H2 cavity (seven missing hole cavity in a PC slab). We also proposed and demonstrated the composite cavity of line and point defects, in which the mode gap of the line defect is used for optical feedback. After that, this type of cavity was optimized, and an ultrahigh Q factor over 10⁶ and a low lasing threshold of 100 mW were reported. In 2004, we obtained the lasing in H1 cavity (one missing hole cavity) with the smallest V at that moment. In 2005, we renewed the smallest record to 0.024 mm³ = 2(l/2n)³ by employing the point-shift cavity with no missing holes (H0 cavity). Such small cavities are not only obtained by the PC but also by the quasi-periodic photonic crystal (QPC). We demonstrated the lasing in H2 and H1 cavities for 12-fold symmetric QPC. The hydrogen-iodide-based ICP etching greatly contributed to this achievement, because it allows the direct etching with resist mask and hence dramatically improves the accuracy of the pattern transfer. In these tiny cavities, the enhancement of the spontaneous emission rate, known as the Purcell effect, is expected. We directly measured the spontaneous decay for the H1 cavity, and confirmed the >16-fold enhancement in 2004. Presently, we are exploring larger enhancement and the thresholdless lasing in the H0 cavity. Besides, we demonstrated the tuning of a PC nanocavity by the carrier injection in 2003. The external control is desired and will be an important issue for PC devices in the near future.

The other important device is the PC waveguide. In these five years, the propagation loss of this waveguide has been reduced to three orders of magnitude lower and reached to a practical level for dense photonic integration. It is time to consider the true uniqueness and effectiveness of this waveguide. A new keyword is the slow light, meaning light with an ultralow group velocity v_g. It is obtained near the photonic band-edge frequency of this waveguide due to the strong structural dispersion. We evaluated v_g ~ c/300 by measuring the mode spacing of the internal Fabry-Perot resonance. However, a narrow bandwidth and the distortion of optical signals due to the dispersion are serious issues for its applications. In 2003, we proposed the chirped PC, in which some structural parameters are gradually changed so that the photonic band is smoothly shifted. It enables us to design the bandwidth. In addition, we proposed a directional coupler and a coupled waveguide for the dispersion compensation. We successfully simulated the slow optical pulse by combining these two concepts. The optical buffer memory by the slow light was selected as a research issue in The Priority Area Research of MEXT (Prof. K. Kobayashi, Tokyo Inst. Tech., as a leader) in 2005. We also investigated the slow light effect in the PC laser and PC optical amplifier. For the former, the slow light effect was clearly shown in the Fabry-Perot lasing mode. For the latter, a high gain was theoretically expected in a 10-mm-long SOA due to the enhanced light-matter interaction. To achieve the current injection and effective heat sinking, all semiconductor PC waveguide with deep holes was explored and the clear light propagation was confirmed. The passive and active integration inside the PC waveguide was also achieved by the regrowth technique through the collaborated research with Furukawa Electric Co., Ltd.

The negative refractive optics based on meta-materials is another emerging field in these three years. However, the absorption loss is a crucial issue for meta-materials at lightwave frequencies. We rather used the anomalous dispersion characteristics of PCs for this purpose, and obtained the clear theoretical simulation of light propagation in the superprism and superlens, for the first time, by optimizing I/O interfaces of the PC in 2004 - 2005. We proposed and experimentally demonstrated the k-vector superprism, which is an alternative to conventional S-vector-based superprism. Here, the refraction of light beam at angled output interface enhances the wavelength resolution. We also investigated unique properties of the superlens, i.e., the real image formation by the flat surface, ultrahigh NA and focusing characteristics insensitive to the input position. The combination of these two elements will realize a compact monochrometer.

In addition to these, we also studied PC LED and PC VCSEL. For the former, we have already showed the enhanced light extraction efficiency in a preliminary experiment in 2002. We further investigated the optimum design for blue/violet devices using full 3D FDTD simulation of incoherent light in a self-made cluster computer, and indicated that the layer structure must be carefully designed to enhance the PC effect. Presently, it is expected to become a practical PC device. The latter is also expected to be a practical device; the single mode output power of VCSELs is improved by the PC. We proposed the triangular holey structure, and achieved the record high power of 7 mW, maintaining a SMSR of 40 dB through the collaborated research with Sony Corp.

Even compared with PCs, high-index-contrast (HIC) structures such as microdisks and wire waveguides are more effective for some applications. For example, the room temperature cw lasing is obtained in microdisks more easily. We achieved a record low lasing threshold of 14 mW for microgear, a modified microdisk cavity, in 2003. The room temperature cw lasing was also obtained in a InAs/AlGaAs quantum-dot microdisk in 2004 – 2005 through the collaborated research with Univ. Tokyo. An important new concept related with the microdisk is the photonic molecule. It is an optical analog to chemical molecules. We clearly observed cavity modal behaviors in the microdisk photonic molecule, which were similar to electronic states in chemical molecules, i.e., the mode splitting, mini-band formation, anti-crossing characteristics, etc., in 2004. Moreover, we successfully demonstrated the bistability and mode switching based on the saturable absorption and nonlinear gain with a low threshold of 70 mW in 2005. Such photonic molecule concept is expected to offer a guideline that explains complex modal behaviors in coupled microcavities.

In these three years, Si photonics has become a topical area in optoelectronics. It is manly because Intel presented Si optical modulators and Raman laser with a novel concept of intra-m-chip optical interconnects. Now, it is believed to be a promising platform for dense photonic integration. In this field, we have already reported sharp bends, branches, intersections, etc. of the Si wire waveguide since 2000, and succeeded in demonstrating a H-tree optical signal distribution circuit in 2002. In 2003 - 2005, we concentrated on demonstrating an ultra-compact arrayed waveguide grating (AWG) demultiplexer and Mach-Zehnder interferometer (MZI) for communication and sensor applications. For the AWG, we confirmed the first operation in 2003, and then improved the performance step by step. Now, the sidelobe level of the AWG is less than -20 dB and the insertion loss is 1.5 dB for a channel spacing of ~10 nm and a device size of (~70 mm)². It will be a practical device if the fiber-pigtailed device is realized with the polarization-insensitive characteristics.

I wish to acknowledge Prof. Y. Kokubun, Prof. H. Arai, and Prof. K. Tada, Yokohama National University, Prof. Y. Suematsu, Prof. K. Iga, Prof. K. Kobayashi, Prof. S. Arai and Prof. F. Koyama, Tokyo Institute of Technology, Prof. S. Kawakami, Tohoku University, Prof. Y. Arakawa, University of Tokyo, Prof. S. Noda, Kyoto University, and Dr. M. Notomi, NTT, for their great support for our research and many valuable suggestions. Also, I would like to thank professors and staffs of Department of Electrical and Computer Engineering, Yokohama National University, those of Precision and Intelligence Laboratory, Tokyo Institute of Technology, Dr. A. Kasukawa, Furukawa Electric, Dr. K. Inada, Fujikura, and PC research members of AIST, Hitachi, NICT, NEC, NIMS, NTT, Ricoh, RIKEN, Sony, etc., for their discussions and supports. Our work was supported by the MEXT, JSPS, JST, and many companies.

April 3rd, 2006

Toshihiko Baba, Professor

(少量，在庫があります．欲しい方は電子メールで連絡してください．)