Oh, Jaesang2011-02-242011-02-242010-12https://hdl.handle.net/11299/100766University of Minnesota Ph.D. dissertation. December 2010. Major: Electrical Engineering, Advisors: Anand Gopinath and James Leger. 1 computer file (PDF); ix, 135 pages. Ill. (some col.)The linearization of the response function of the optical directional coupler modulator uses variable coupling along the length of the device, and the coupling function can be synthesized for a desired response function. This synthesis gives rise to coupling functions which have negative coupling regions realized by placing a phase shift in one of the arms of the coupler, and the reversion to the positive coupling requires a second phase shift in one or other arm of the coupler. A linear response modulator was previously designed and fabricated, but obtained large switching voltages, and an alternative trapezoidal response was chosen. A second design using a uniform coupling directional coupler with four ð phase shifts to obtain a linear response was considered. These two linear designs have been investigated, the trapezoidal response's design was synthesized by the Fourier transform technique and the design was refined using the Newton's method. The phase shift design was achieved by four phase shift points, which were chosen very carefully for linear response. The small signal calculations for the second and third order intermediate distortions (IMD2 and IMD3) show Spur Free Dynamic Ranges (SFDR) of the better than 120 dB/Hz2/3 for the IMD3 for both designs, and the IMD2 of the order of 105 dB/Hz1/2 for the trapezoidal response and better than 120 dB/Hz1/2 for the phase shift design. The switching voltage-length product of the trapezoidal response is lower than the phase shift design. Since the electrical phase velocity in III-V semiconductor coplanar strip electrode structures is higher than the optical phase velocity, a slow-wave electrode structure is required for velocity-matching, and the T-rail shunt capacitive and series inductive loaded coplanar strip was chosen. These optical directional modulator designs have been fabricated and tested. The intensity responses are different from the design because of fabrication problems including uneven waveguide etching along the device. The linearity measurement was not performed because of the weak optical output from the modulator due to high insertion loss of the modulator. The velocity matched designs cover up to 20 GHz and is limited by the measurement equipment.en-USIII-V Semiconductor MaterialIterative Newton's MethodLinear Optical ModulatorNegative CouplingSFDRSlow-waveElectrode StructureElectrical EngineeringThesis or Dissertation