The explosive growth of data-centric artificial intelligence applications calls for the next generation of optical interconnects for future hyperscale data centers and high-performance computing (HPC) systems. To unleash the full potential of dense wavelength-division multiplexing, we present the design and exploration of a novel transceiver architecture based on silicon photonic micro-resonators featuring a broadband Kerr frequency comb source and fabrication-robust (de-)interleaving structures. In contrast to the traditional single-bus architecture, our architecture de-interleaves the comb onto multiple buses before traversing separate banks of cascaded resonant modulators/filters, effectively doubling the channel spacing with each stage of de-interleaving. With a closed-form free spectral range (FSR) engineering technique guiding the micro-resonator design, the architecture is scalable toward hundreds of parallel channels—spanning much wider than the resonator FSRs—with minimal crosstalk penalty and thermal tuning overhead. This unique architecture, designed with co-packageability in mind, thus enables a multi-Tbps aggregated data rate with moderate per-channel data rates, paving the way for sub-pJ/b ultra-high-bandwidth chip-to-chip connectivity in future data centers and HPC systems.
While the high index contrast between silicon and silicon dioxide in the silicon-on-insulator photonics platform permits unprecedented device density, it also leads to high sensitivity to fabrication variations. In silicon microring and microdisk resonator devices, fabrication variations can substantially change the target resonance wavelength. Silicon’s high thermo-optic coefficient allows for correction of these fabrication variations and stabilization of the device resonant wavelength through thermal tuning. Metal and doped silicon integrated heaters are commonly used to perform this tuning and have become an essential feature of silicon microring and microdisk modulators. Metal heaters are typically placed in a layer above the silicon devices, while doped silicon heaters are placed in the same silicon waveguide layer, adjacent to the devices. The advantage of doped silicon heaters over metal heaters is due to proximity of the heater to the optical device, leading to greater efficiencies. However, for active devices using p-n junctions such as modulators, parasitic junctions can form between the doped heater and the modulator junctions, resulting in highly unstable and substandard device performance. Here, we present a detailed simulation framework for heater design in resonant silicon microdisk modulators, supported by experimentally measured device performance, which emphasizes tuning efficiency while eliminating parasitic diode formation. Simulations were conducted in Ansys Lumerical HEAT, CHARGE, and MODE to model parasitic junction behavior between the heater and modulator, in addition to the heater’s thermal response and its effect on the resonant wavelength of the microdisk.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.