L.E. Johnson, H. Xu, D.L. Elder, S.R. Hammond, S.J. Benight, B.H. Robinson
Nonlinear Materials Corporation,
United States
Keywords: hybrid photonics, electro-optics, organic materials, nonlinear optics, optical computing, organic semiconductors, materials modeling
Summary:
Computing and telecommunications are presently undergoing a massive transformation. Semiconductor circuit density (Moore’s law) is approaching physical limits, computing is consuming a rapidly growing fraction of the world’s electricity supply, and computing trends such as cloud and edge computing require large interconnect bandwidths at distances ranging from long-distance optical networks to between modules within individual chips. Optical processing has also recently become of great interest for machine learning/neural network applications. Cost-efficient implementation requires components at similar scales to electronic components and materials that can be integrated with conventional silicon CMOS technologies. Organic materials have previously been considered for high-speed electro-optic technologies due to exceptional electro-optic activity and THz intrinsic bandwidth, but were limited by the need for polymer binders, high optical loss, and limited device lifetimes. Hybrid device architectures, such as plasmonic-organic hybrid (POH) and silicon-organic hybrid (SOH) photonics systems present a promising approach that leverages the best aspects of organic electro-optic (OEO) materials and conventional semiconductor manufacturing processes. The refractive index of the organic layer in hybrid devices is incredibly sensitive to electric fields ( > 20X the electro-optic response of standard EO material lithium niobate), enabling signal detection and modulation within micro/nano-scale devices. Integration with silicon or plasmonic systems provides a platform for deployment of standard nanofabrication techniques and integration with electronics. Plasmonic components can further concentrate electric fields into the EO material, enabling smaller and more sensitive devices. Smaller device sizes also minimize issues with optical loss in the organic layer, allowing use of materials with stronger EO activity than in bulk devices. Research milestones realized using previous generation materials include > 500 GHz bandwidth in modulator applications, power consumption on the order of 70 aJ/bit, and device footprints < 10 μm2. We have developed and applied computational tools for designing future generations of electro optic materials with unprecedented performance and applied these tools towards development of multiple families of novel OEO materials. Recent developments include materials that combine high thermal stability with large electro-optic activity and materials capable of realizing electro-optic activities on the order of 1000 pm/V