Semiconductor Mach-Zehnder (MZ) modulators and arrayed waveguide gratings (AWGs) generally use deep-ridge waveguides to achieve a compact device circuit. However, deep ridge waveguides are easy to induce higher order lateral modes. Simply designing these higher order modes as leaky modes is not enough because they may still propagate to the end of the device circuit and degrade the device performance if the leaky rate is not fast enough. Therefore these high order leaky modes should be well analyzed and controlled in practice. We combine the compact 2D finite-difference time-domain (FDTD) technique with the Padé approximation transform to analyze these leaky higher order modes.
The same method is also used to design 40G GaAs phase modulators.
Current work: Design of Novel Lasers with Slots
Single mode and widely tunable semiconductor lasers are widely used in optical communications.
Typically these lasers are based on low perturbation buried gratings (DFB or DBR).
Fabricating these lasers generally requires both high resolution processing and complex re-growth steps.
Single mode and tunable lasers can also be realized by simply introducing distributed reflective defects (slots)
into the ridge of conventional ridge waveguide Fabry-Pérot (FP) laser. We use the 2D scattering matrix method (SMM)
to analyze the properties of these slots.
Laser threshold as low as 10 mA and SMSR up to 51 dB for a slotted single mode laser suitable for photonic integration has been predicted.
Figure 4.(a) 3D schematic structure of the slotted single mode laser. (b) Simulated output power and SMSR versus injected current for the optimized slotted laser structure. (c) Simulated SMSR versus the cavity mode wavelength shift caused by the uncertainty of the cleave facet position with the current injection of 100mA is assumed in the simulation.
Discretely tunable multi-section tunable laser built on slots is also demonstrated.
Figure 5.(a) Calculated power reflection spectrum for a 3 section tunable laser based on multi-slots with channel spacing of around 400 GHz (3.2 nm). (b) Simulated lasing wavelength versus currents in mirror sections. (c) Simulated SMSR versus wavelength.