Polymer promises wideband modulators

A nonlinear optical-polymer-based electro-optic modulator with a bandwidth of 60 GHz has been demonstrated by a group of researchers from the University of California (UCLA, Los Angeles, CA) and the University of Southern California (USC, Los Angeles, CA). This bandwidth is about three times greater than the typical bandwidth of devices made from lithium niobate (LiNbO₃). The researchers are also evaluating an extension of the technology to 95 GHz using new electrode designs. According to Larry R. Dalton, professor of chemistry and electrical engineering at USC, it is the signal loss at the modulator`s metal electrodes that limits the operating bandwidth of the device.

For optoelectronic applications, organic second-order nonlinear optical polymers offer several putative advantages over the alternative crystalline nonlinear materials. The lattice constant mismatch between crystalline electro-optic materials, such as LiNbO₃, and the electronic materials, such as silicon and gallium arsenide, means that monolithic integration of these technologies is difficult and potentially expensive. Polymeric devices, on the other hand, are potentially lower cost with optical and electrical properties that can be tailored to suit a specific application. They have a low dielectric constant, which is useful for high-speed modulation, and lend themselves to monolithic integration; they can also be relatively easily combined with silicon-based VLSI (very large scale integration) circuitry.

To be useful as a modulator, an optical polymer must contain highly polarizable chromophores--to provide a high degree of nonlinearity--and be thermally stable enough to withstand the harsh processing conditions encountered during optoelectronic device fabrication. Nonlinear optical activity occurs when all the chromophore branches on the polymer backbone are lined up parallel to each other and fixed. Various techniques have been developed in order to fabricate materials that meet these conditions. These methods include the use of electric fields and lasers to align the molecules (poling), but the resulting polymer has typically been unstable once the poling field is removed.

The UCLA/USC researchers ad dressed this problem by developing a novel fabrication process that permits highly active chromophores to be incorporated at high concentrations into polymer materials such that subsequent chromophore decay is minimized. Fabrication, explains Dalton, is based on preparation of functionalized nonlinear optical chromophores that are used to make processable precursor polymers, characterized by modest glass-transition temperatures. The precursor polymer is then combined with conventional solvents to spin-cast optical-quality films. An external poling field orders the chromophores before the polymer is finally hardened by condensation into robust lattices.

Also, says Dalton, unlike competing materials technology that involves incorporation of chromophores into polymers with high glass-transition temperatures, such as polyimides, this precursor approach to making polymers is compatible with standard VLSI clean-room processing. This approach also yields materials that are amenable to fabrication of buried-channel waveguides and for integration with other system components. The electro-optic coefficients of the precursor materials are in the 15 to 50 pm/V range, and thermal stability is such that 95% of the optical nonlinearity is retained at 100°C for more than 1000 h.

As development of stable nonlinear optical polymers progresses, researchers are also actively investigating integration of electronic and optical devices. The UCLA/USC team has recently demonstrated the feasibility of fabricating active-polymer modulators on VLSI circuits with no degradation of the performance of either the optical or electronic components. Before full integration of working VLSI circuits and polymeric devices is achieved, however, issues such as interconnects and access vias through the various layers must be addressed.