Ultrahigh-speed on-chip all-optical amplitude modulation via monolayer graphene saturable absorption

by Michael Thomas Hoff

Institution: UCLA
Year: 2016
Keywords: Electrical engineering; all-optical modulation; graphene; material optics; on-chip; photonics; saturable absorption
Posted: 02/05/2017
Record ID: 2090049
Full text PDF: http://www.escholarship.org/uc/item/34j9r5x0


As global internet traffic continues to climb at an uncurbed rate, the development of correspondingly fast physical communications technology is as crucial as ever. Key milestones have already been reached with graphene-integrated waveguide structures in the field of electro-optic modulators, bolstering hope that practical ultrahigh-speed graphene optical modulators could indeed be actualized, with the potential to yield revolutionary results. But while the electrical gate-tunability of these devices has been soundly demonstrated, there remains the question of modulation speed, limited primarily by the parasitic RC-time intrinsic to such electrical control. Recent advancements have already shown graphene's robust performance as a broadband saturable absorber, due to its massless Dirac fermions and linear dispersion relation near the Dirac point. By capitalizing on these material properties and moving away from electrical control into the all-optical regime, the only limitation on device operation speed is due to the intrinsic material properties of graphene – namely, excitation, thermalization, and recombination times of the excited fermionic carriers. In this work, an on-chip microring resonator amplitude modulator is constructed, and a theoretical analysis of the intrinsic and extrinsic speed limitations on its all-optical operation is developed. Laboratory data is presented, and a computational model is built to further enhance an understanding of the modulator's operation and potential limitations. It is found that laboratory data confirms material response times corresponding to modulation speeds exceeding 50 GHz – although limited by the experimental instrumentation rather than the device itself – while the numerical model reveals a higher performance threshold and suggests optimizations to achieve faster speeds in the near future.