For the SOI-waveguide directional coupler (WDC), optical access loss (OAL) and polarization dependence (PD) are two
critical performance specifications which seriously affect the adoptability and deployment of a device, including optical
on-chip loss (OCL), polarization dependent loss (PDL) and extinction ratio of a 3dB-coupler based device. In this work,
using a commercial software tool - FIMMPROP, the performance of an SOI-WDC is simulated. Simulations find that the
curved waveguides for the turning sections of a 3dB WDC not only enlarge the footprint size, but also seriously
deteriorate the device performance. For instance, the two curved waveguide sections of a WDC induce an unpredictably
large change in the 3dB-coupling length, increase an OAL of 0.4-0.9dB, and seriously deteriorate the PD, and these
performance changes radically depend on rib size. After a corner-turning mirror (CTM) structure is introduced to a 3dB
SOI-WDC, the experiments show both the footprint length and 3dB-coupling length are unchanged, the OAL of the 3dB
coupler is only 0.5dB which is close to the simulation value. Therefore, for a 3dB-coupler based Mach-Zehnder
interference (MZI) structure, the OCL will be controlled to be <1.0dB in device design and will not depend on rib size.
SiO2-TiO2 planar optical waveguides are fabricated on silicon wafer substrate by dip-coating technique with the
Sol-Gel solutions, based on which the stripe optical waveguides are patterned by laser direct writing of the Sol-Gel films
using an Ytterbium fiber laser and followed by chemical etching. The effects of the laser processing parameters on the
microstructure of the core layer films are investigated. The relative chemical etching rates of the non-irradiated area in
Sol-Gel films that are haeted at different temperature are characterized. The optical fields and propagation losses of the
optical waveguides at the wavelength of 1550 nm are characterized by multi-channel fiber/waveguides coupling system.
The experimental results demonstrate that the composition, the post heat treatment temperature and laser power density
have a big effect on the widths of the stripe optical waveguides, and the minimum widths about 25 μm can be fabricated
with the suitable parameters. The core layer of the planar optical waveguides as received by Sol-Gel method is loose in
structure, and a shrinkage concave groove forms in the laser irradiated area. The microstructure and forming mechanisms
of the stripe waveguides by laser direct writing Sol-Gel films are discussed. The minimum propagation loss of the
fabricated stripe waveguides is about 1.77dB/cm at 1550nm. Better results are expected by improving the film
composition and laser processing parameters further.
A thermo-optic variable optical attenuator (VOA) based on a Mach-Zehnder interferometer and multimode-interference coupler is fabricated. Not a single-mode but a multimode waveguide is used as the input and output structures of the optical field, which greatly reduces the coupling loss of the VOA with a normal single-mode fiber. The insertion loss of the fabricated VOA is 2.52 to 2.82 dB at the wavelength of 1520 to 1570 nm. The polarization dependent loss is 0.28 to 0.45 dB at the same wavelength range. Its maximum attenuation range is up to 26.3 dB when its power consumption is 369 mW. The response frequency of the fabricated VOA is about 10 kHz.
Large primary mirrors for optical telescope are usually interferometrically tested using null correctors. In fabrication of primary mirror, the optical surface is polished to precisely match the wavefront generated by the null corrector. The final shape of the primary mirror will be incorrect if a flawed null corrector has been used.In this paper a new test to certify null corrector was designed and implemented that uses small computer—generated holograms (CGH) fabricated onto flat substrates. This test solves the difficult problem of verifying the accuracy of the null correctors. A new technique for hologram fabrication has been explored for this application. A laser writing machine was built for fabricating these patterns onto flat substrates. The null corrector tests the hologram exactly as if a real mirror is to be measured. The use of hologram to test the null corrector is surprisingly simple. The CGH is positioned at the paraxial focus ofthe mull corrector. Once the CGH is near the correct position, the shape of the fringes in the interferometer is used to align the hologram. Since the CGH appears to the null corrector to be a complete primary mirror with correct shape, the corrector can be tested exactly.