There can be no commercial poled-polymer optical waveguide switches until several goals are reached. Among these are thermal stability, low switching voltage, and low-loss waveguides with acceptable fiber insertion loss. We report on two cross-linked materials; one shows r33 values above 10 pm/V and retains 80% of its E-O coefficient after 1500 hours at 100 degree(s)C. Using these systems for illustration, we discuss material requirements and design tradeoffs to optimize performance.
Directional couplers based on titanium-diffused lithium niobate are potentially attractive for applications requiring polarization-independent optical switching or modulation. Although directional couplers with >25 dB switching extinction ratio for both TE and TM modes have been fabricated, polarization independence and reproducibility are difficult to achieve. Presumably, uncontrollable process variations result in different coupling strengths and switching characteristics in nominally identical devices. These variations can be compensated by using the so-called reverse-(Delta) (beta) design, where the device electrodes are split into two different sections and tuned with separate voltages to obtain maximum cross-state extinction. The addition of a second control voltage complicates characterization of the devices, since full characterization requires plotting switching in a two-dimensional voltage space. The authors have developed a system to automatically perform two-dimensional switch characterizations. The resulting data display asymmetries which can be analyzed within the framework of coupled mode theory. The analyses indicate that two factors contribute to the observed asymmetry: a static, fabrication-induced difference in propagation constant (beta) between the two waveguides, and an applied-field dependent evanscent coupling strength.
The dynamical properties of LiNbO3 switches were measured as a function of radiation. The devices are titanium indiffused delta beta switches. Measurements of the gamma radiation level using radiation monitors indicate a total dose of less than 100 rad for a 5-min gamma exposure. The dc levels attenuate for the unswitched case, but not to the same degree that the switched signals do. With pulsed beam, a pulsed attenuation of the signal is evident.