Since the turn of the century, Silicon Photonics (SiPh) has advanced data communication and computing via diverse integrated optical functions and multiplexing strategies. However, conventional design methodologies limit scalability, and inverse designs lead to features sensitive to fabrication process variations. This talk explores the harnessing of machine learning (ML) to predict and rectify these deviations in the design phase. This technique enhances design fidelity and device performance, while facilitating smaller design features, bypassing constraints of traditional methods. Highlighting PreFab, an innovative ML technology, applicable to both conventional and inverse designs, it predicts and corrects fabrication deviations, enabling refined design processes.
Recent developments in computational inverse design offer the promise of significantly reducing the footprint and allowing complex optical functionality in silicon photonic components as compared to existing conventional building blocks. However, reliable fabrication of such components is one of the major bottlenecks in its widespread adoption. A common characteristic of such designs is the presence of small features that have meaningful impact on the optical performance. Current approaches to tackle this problem consider designing for robustness, such as by co-optimizing for over- and under-etched geometries at the design stage. This is often followed with a design-for-manufacturing optimization step to meet specifications of a foundry such as minimum feature size and curvature radii. Those approaches often incur additional significant computational costs as well as a reduction in peak optical performance. In this work, we highlight our recent progress to bridge the gap between inverse design methods and their ability to deliver reliable and manufacturable designs. We observe that the so-called parameterized shape optimization methods are more likely to produce robust designs for certain classes of components, as showcased in integrated mode converter designs. For components that benefit from topological inverse design such as wavelength demultiplexers, we propose a new optimization penalty that naturally leads the optimizer towards more robust designs. In a new research direction, we also consider improving fabrication reliability by the development and use of data-driven predictive models for fabrication. Leveraging deep learning tools, we present prediction and correction models that improve fabrication outcomes for a variety of components made at an e-beam prototyping foundry.
For next-generation on-chip optical communications, the combination of wavelength- and mode-division multi-plexing has been proposed as way to achieve high aggregate bandwidth on a single multimode silicon bus waveguide. In this application, microring resonators are used to multiplex and demultiplex the different wavelength and mode channels. The designs of the microrings ultimately determine the throughput and power consumption of the circuit. In this work, we perform a comprehensive numerical comparison of the two designs that have already been demonstrated (the point-coupled and racetrack microring resonators) with a new, waveguide-wrapped design. The proposed design reduces the size of the microring cavity, which reduces the spatial footprint and increases the free spectral range (bandwidth), without heavily increasing loss and intermodal crosstalk (power budget). We believe that efficiency improvements to next-generation optical communications circuits like these can help meet the growing global demand for bandwidth without adding to the growing cost of operating them.
We present a theoretical model for two high-throughput optical logic methodologies, using voltage-induced free-carrier dispersion and stimulated Raman scattering based Zeno switching. Increased computational throughput is achieved by accessing higher switching speeds, optimizing the use of space, and by using multiple wavelengths for parallel processing. The condition of CMOS compatibility is maintained to take advantage of the high-volume, low-cost manufacturing potential of the industry and to help lower each design's spatial footprint (enabled by the high refractive index contrast of silicon-on-insulator waveguides and resonators). Each design is made with the potential of higher-order operations in mind; for their use must not only stand alone, but must also have the ability to incorporate into future all-optical or optoelectronic computational devices.
For decades, the semiconductor industry has been steadily shrinking transistor sizes to fit more performance into a single silicon-based integrated chip. This technology has become the driving force for advances in education, transportation, and health, among others. However, transistor sizes are quickly approaching their physical limits (channel lengths are now only a few silicon atoms in length), and Moore's law will likely soon be brought to a stand-still despite many unique attempts to keep it going (FinFETs, high-k dielectrics, etc.). This technology must then be pushed further by exploring (almost) entirely new methodologies. Given the explosive growth of optical-based long-haul telecommunications, we look to apply the use of high-speed optics as a substitute to the digital model; where slow, lossy, and noisy metal interconnections act as a major bottleneck to performance. We combine the (nonlinear) optical Kerr effect with a single add-drop microring resonator to perform the fundamental AND-XOR logical operations of a half adder, by all-optical means. This process is also applied to subtraction, higher-order addition, and the realization of an all-optical arithmetic logic unit (ALU). The rings use hydrogenated amorphous silicon as a material with superior nonlinear properties to crystalline silicon, while still maintaining CMOS-compatibility and the many benefits that come with it (low cost, ease of fabrication, etc.). Our method allows for multi-gigabit-per-second data rates while maintaining simplicity and spatial minimalism in design for high-capacity manufacturing potential.
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.