SystemX

For over half a century, Integrated Circuits have been designed and developed (rather successfully) toward the goal of enhancing computing performance and efficiency. During this time, the relationship between circuit design and computing has remained largely one-directional: Careful, detailed circuit design is performed in the service of building computing systems. Notwithstanding a post-Moore and post-Dennard reality, the impressive strides made by digital computing thus far prompt an important question which re-examines the traditional circuit-computing relationship: Can runtime computing itself be used to enhance circuit and system capabilities? If so, under which conditions and to what extent?

In this talk, I will present recent efforts in my group that represent two different ways that computing can augment circuit capabilities to (1) overcome limitations inherent in circuit design; and (2) enable rapid, time-optimal control of integrated control systems. The effectiveness and limitations of both efforts are examined through a representative test-chip design. These efforts have yielded a robust True-Random Number Generators (TRNGs) demonstrating the lowest measured energy-per-bit (2.58pJ/bit), and an all-digital PLLs (ADPLLs) for system clocking applications with the fastest demonstrated cold-start and re-lock times (16 Refclk cycles, mean).

We develop miniaturized integrated bio/analytical instruments and platforms to conduct economical, frequent, autonomous life-science experiments in outer space. The technologies represented by several of our recent 5-kg "free-flyer" small-satellite missions are the basis of a rapidly growing suite of miniaturized biologically- and chemically-oriented instrumentation now enabling a new generation of in-situ space science experiments. Over the past decade, our missions have included studies of space-environment-related changes in gene expression, drug dose response, microbial longevity and metabolism, and the degradation of biomarker molecules. The science and technology of one of these missions, the O/OREOS (Organism/Organic Response to Orbital Stress) Nanosatellite, will be highlighted in the context of conducting biological and chemical experiments in outer space using miniaturized integrated systems.

We have recently begun to adapt and apply our spaceflight-compatible microfluidic and bioanalytical technologies to the challenge of finding molecular and structural indications of microbial life on the so-called icy worlds of our solar system, particularly the moons Europa and Enceladus. The design, development, and laboratory testing of the Sample Processor for Life on Icy Worlds (SPLIce) system, a microfluidic sample-processing "front end" to enable autonomous detection of signatures of life and measurements of habitability parameters on icy worlds, will be described. SPLIce is under development to support two nominal mission scenarios: a fly-through of Enceladus' icy plumes, expected to yield ~ 2 µL of ice particles/square meter of collector area, and a Europan lander, the rasp-based sampling system of which is anticipated to deliver 1 – 5 mL of icy solids for analysis.

Individuals of all genders invited to be a part of:
#StanfordToo: A Conversation about Sexual Harassment in Our Academic Spaces, where we will feature real stories of harassment at Stanford academic STEM in a conversation with Provost Drell, Dean Minor (SoM), and Dean Graham (SE3). We will have plenty of time for audience discussion on how we can take concrete action to dismantle this culture and actively work towards a more inclusive Stanford for everyone. While our emphasis is on STEM fields, we welcome and encourage participation from students, postdocs, staff, and faculty of all academic disciplines and backgrounds.

Quantum Computing has been getting lots of buzz recently but does its current state live up to the hype? This talk will survey the state of the industry from the perspective of a venture capitalist.

Next-generation nanoelectronics for logic and memory are based on devices increasingly smaller, more three-dimensional in shape and containing even more types of materials. Evaluating nanometre-scale features of such devices, including carrier profiling, strain, electrical and chemical properties represents a challenge for materials catherization. Here, after introducing the current and proposed logic and memory devices at advanced nodes, I will present dedicated two- and three-dimensional analysis methods to merge our present capability of materials characterization with site-specific and failure analysis. Different techniques are presented and combined to maximize our sensing capability and boost the process development of various emerging technologies, including fin-based field-effect transistors (FinFET) and various types of non-volatile resistive switching memories.

Computing near the sensor is preferred over the cloud due to privacy and/or latency concerns for a wide range of applications including robotics/drones, self-driving cars, smart Internet of Things, and portable/wearable electronics. However, at the sensor there are often stringent constraints on energy consumption and cost in addition to the throughput and accuracy requirements of the application. In this talk, we will describe how joint algorithm and hardware design can be used to reduce energy consumption while delivering real-time and robust performance for applications including deep learning, computer vision, autonomous navigation/exploration and video/image processing. We will show how energy-efficient techniques that exploit correlation and sparsity to reduce compute, data movement and storage costs can be applied to various tasks including image classification, depth estimation, super-resolution, localization and mapping.

The use of unconventional devices, along with exploitation of unique properties of the III-N material system, has the potential to significantly impact both high frequency and high-power systems. In this talk, several device technologies being explored at the University of Notre Dame will be described. Recent demonstrations of scaled GaN-based HEMTs having experimentally-demonstrated ft's of over 370 GHz, indicate that GaN-based devices are attractive not just for microwave power amplification, but also for mm-wave and mixed-signal circuit applications as well. The aggressive device scaling that enables these high speeds also led to the observation of room-temperature plasma wave propagation in GaN-channel HEMTs, suggesting additional avenues for high-frequency device design as well as phenomena to be leveraged for enhanced functionality. In addition, the large polar optical phonon energy in GaN results in transport properties that are distinctly different from most other III-V materials; this can be exploited to engineer devices for improved performance. The details of impact ionization in GaN—important for both power devices and devices such as IMPATT diodes—have also recently been explored and will be presented. Novel processing schemes for heterogeneous integration, high-field edge termination, and thermal management of III-N devices—critical to the exploiation of GaN for high power applications—will also be presented.

While lithography was a key technology for making transistors into complex integrated circuits, micro-assembly is potentially a key technology for making nanotechnology into large, complex, heterogeneous, custom systems. We aim to build a new tool for integrating millions of pre-fabricated chiplets or micro-objects into systems, based on deterministic micro-assembly and transfer. The process uses chips initially in solution, and then sorts, transports, and orients chips with directed electrostatic assembly and parallel control. Assemblies are then transferred to final substrates with a stamp or continuous feed roll-based methods, and then electrically interconnected. The current laboratory systems have handled small chips (10 um – 500 um), demonstrated fine registration (<1 um and <1°), and produced centimeter scale outputs. Ultimately, massively parallel automated microassembly, analogous to a xerographic printer using microchips instead of toner, could be used for integrating circuits, microLEDs and other semiconductor components into complex, heterogeneous systems.

Quantum computing is a paradigm for computing that potentially allows certain classes of problems to be solved far more quickly than is possible on conventional computers. There has recently been large progress in the development of intermediate-scale quantum-computer prototypes, thanks to commitments from a number of industrial players, including IBM, Google, Microsoft, Intel, and several startup companies. In this talk I will introduce what quantum computers are, what promise they have to solve difficult computational problems, what problems we can expect might be solved using upcoming hardware prototypes over the next few years, and what major outstanding challenges for the field remain. I will emphasize problems in optimization, quantum simulation, and machine learning.

Neural implants need to establish stable and reliable interfaces to the target structure for chronic application in neurosciences as well as in clinical applications. They have to record electrical neural signals, excite neural cells or fibers by means of electrical stimulation. In case of optogenetic experiments, optical stimulation by integrated light sources or waveguides must be integrated on implants. Metabolic monitoring and detection of neurotransmitter concentrations is also part of the research agenda but not yet mature enough for translation in chronic clinical applications. Proper selection of substrate, insulation and electrode materials is of utmost importance to bring the interface in close contact with the neural target structures, minimize foreign body reaction after implantation and maintain functionality over the complete implantation period. Our work has focused on polymer substrates with integrated thin-film metallization as core of our flexible neural interfaces approach and silicone rubber with metal sheets. Micromachining and laser structuring are the main technologies for electrode array manufacturing. Designing applications for implants in the peripheral and central nervous system needs integration of components, the connection of cables and connectors to both, electrode arrays and hermetic packages containing electronic circuitry for recording, stimulation and signal processing. Failure of one of the components or connections stops the function of the whole system. We present an exemplary implant system and discuss state of the art materials and manufacturing techniques as well as prominent failure modes. Thin-film substrates and hybrid combinations with silicone rubber substrates serve as neural interfaces. Adhesion layers have been integrated to obtain long term stability of polyimide-platinum sandwiches. Hermetic packages with dozens of electrical feed-throughs need novel approaches to meet the desire of implants with hundreds of electrode channels. Reliability data from long-term ageing studies and chronic experiments show the applicability of thin-film implants for stimulation and recording and ceramic packages for electronics protection. Examples of sensory feedback after amputation trauma, vagal nerve stimulation to treat hypertension and chronic recordings from the brain surface display opportunities and challenges of these miniaturized implants. System assembly and interfacing microsystems to robust cables and connectors still is a major challenge in translational research and transition of research results into medical products.