Kyunglok Kim

Optimized Circuit Failure Prediction for Aging: Practicality and Promise

Circuit failure prediction is used to predict occurrences of circuit failures, during system operation, before errors appear in system data and states. This technique is applicable for overcoming major scaled-CMOS reliability challenges posed by aging mechanisms such as Negative-Bias-Temperature-Instability (NBTI). This is possible because of the gradual nature of degradation associated with such aging mechanisms. Circuit failure prediction uses special on-chip circuits called aging sensors. In this paper, we experimentally demonstrate correct functionality and practicality of two flavors of flip-flop designs with built-in aging sensors using 90 nm test chips. We also present an aging-aware timing analysis technique to strategically place such flip-flops with built-in aging sensors at selective locations inside a chip for effective circuit failure prediction. This aging-aware timing analysis approach also minimizes the chip-level area impact of such aging sensors. Results from two 90 nm designs demonstrate the practicality and effectiveness of optimized circuit failure prediction with overall chip-level area impact of 2.5% and 0.6%.

Gate-Oxide Early Life Failure Prediction

This paper uses 90nm transistor-level experimental data, device modeling, and circuit simulations to establish the following results: 1. A transistor with defective gate- oxide, i.e., a gate-oxide early-life failure (ELF) candidate transistor, produces gradually degraded drive currents over time before it completely loses its transistor characteristics; 2. The above phenomenon results in gradual increase in delays of digital circuit paths containing the ELF candidate transistor before the circuit produces functional failures; 3. Gradual delay shifts caused by ELF candidate transistors are large enough to be detected using inexpensive digital techniques. These results can be utilized to overcome scaled-CMOS reliability challenges through ELF identification during production test or on-line during system operation.

Bio-inspired stretchable network-based intelligent composites

The human skin hosts an array of sensors that are capable of detecting and interpreting many traits important to how we function and survive. The goal of mimicking this capability in composites to create intelligent composite materials has led to the development of a bio-inspired stretchable network composed of numerous micro-fabricated sensors capable of detecting multiple stimuli. The components of the network are small scale and flexible making the network embeddable within complexly shaped composite layups and flexible structures with minimal impact on the host structure. This paper outlines recent progress in ongoing work to develop the bio-inspired network in order to create intelligent composite materials.

Experimental study of gate oxide early-life failures

Large-scale experimental data from 90nm test chips consisting of 49,152 transistors, and experiments on 90nm test chips containing inverter chains are used to establish: 1. A gate-oxide early-life failure (ELF, also called infant mortality) candidate transistor produces gradually degraded drive currents over time; 2. A digital circuit path consisting of a gate-oxide ELF candidate transistor experiences gradual delay shifts over time before the circuit produces functional failures. These results may be utilized to effectively overcome ELF challenges in scaled CMOS technologies.

Micro-fabricated, expandable temperature sensor network for macro-scale deployment in composite structures

We have developed methods for creating a highly expandable temperature sensor network for distributed temperature measurement. Stresses and strains due to network expansion are minimized through finite element analysis. Through the use of a uniquely patterned polyimide substrate and wire pattern an expansion ratio of 1,000% is achieved and the electrical resistance of components is maintained from pre-expansion to full expansion. Platinum resistance temperature detectors and electrodes are integrated directly in the polyimide-based network through a non-standard micro fabrication process. Calibration and interpolation algorithms have been developed for temperature measurement. Real-time distributed temperature measurement has been achieved through this sensor network, and it has shown great potential to be integrated into composites. 1 2

High performance wash-free magnetic bioassays through microfluidically enhanced particle specificity

Magnetic biosensors have emerged as a sensitive and versatile platform for high performance medical diagnostics. These magnetic biosensors require well-tailored magnetic particles as detection probes, which need to give rise to a large and specific biological signal while showing very low nonspecific binding. This is especially important in wash-free bioassay protocols, which do not require removal of particles before measurement, often a necessity in point of care diagnostics. Here we show that magnetic interactions between magnetic particles and magnetized sensors dramatically impact particle transport and magnetic adhesion to the sensor surfaces. We investigate the dynamics of magnetic particles’ biomolecular binding and magnetic adhesion to the sensor surface using microfluidic experiments. We elucidate how flow forces can inhibit magnetic adhesion, greatly diminishing or even eliminating nonspecific signals in wash-free magnetic bioassays, and enhancing signal to noise ratios by several orders of magnitude. Our method is useful for selecting and optimizing magnetic particles for a wide range of magnetic sensor platforms.

Bio-Inspired Stretchable Absolute Pressure Sensor Network

A bio-inspired absolute pressure sensor network has been developed. Absolute pressure sensors, distributed on multiple silicon islands, are connected as a network by stretchable polyimide wires. This sensor network, made on a 4’’ wafer, has 77 nodes and can be mounted on various curved surfaces to cover an area up to 0.64 m × 0.64 m, which is 100 times larger than its original size. Due to Micro Electro-Mechanical system (MEMS) surface micromachining technology, ultrathin sensing nodes can be realized with thicknesses of less than 100 µm. Additionally, good linearity and high sensitivity (~14 mV/V/bar) have been achieved. Since the MEMS sensor process has also been well integrated with a flexible polymer substrate process, the entire sensor network can be fabricated in a time-efficient and cost-effective manner. Moreover, an accurate pressure contour can be obtained from the sensor network. Therefore, this absolute pressure sensor network holds significant promise for smart vehicle applications, especially for unmanned aerial vehicles.

High-Resolution Analysis of Antibodies to Post-Translational Modifications Using Peptide Nanosensor Microarrays

Autoantibodies are a hallmark of autoimmune diseases such as lupus and have the potential to be used as biomarkers for diverse diseases, including immunodeficiency, infectious disease, and cancer. More precise detection of antibodies to specific targets is needed to improve diagnosis of such diseases. Here, we report the development of reusable peptide microarrays, based on giant magnetoresistive (GMR) nanosensors optimized for sensitively detecting magnetic nanoparticle labels, for the detection of antibodies with a resolution of a single post-translationally modified amino acid. We have also developed a chemical regeneration scheme to perform multiplex assays with a high level of reproducibility, resulting in greatly reduced experimental costs. In addition, we show that peptides synthesized directly on the nanosensors are approximately two times more sensitive than directly spotted peptides. Reusable peptide nanosensor microarrays enable precise detection of autoantibodies with high resolution and sensitivity and show promise for investigating antibody-mediated immune responses to autoantigens, vaccines, and pathogen-derived antigens as well as other fundamental peptide-protein interactions.

Functionalization of stretchable networks with sensors and switches for composite materials

An investigation was performed to develop appropriate techniques to design and fabricate (using complementary metal-oxide semiconductor/micro-electro-mechanical systems technologies) highly stretchable networks of distributed sensors and organic diodes that could be stretched, and surface-mounted or embedded into polymeric materials to cover an area several orders of magnitude larger than its original size. Both analysis and experiments were performed to validate the design and fabrication methods. The techniques sought to reduce stresses due to network expansion, and a new spin-coated fabrication process was developed to enable high-resolution features in the network. Networks with temperature sensors and piezoelectric sensors were fabricated and tested to demonstrate functionality in advanced composite materials that are common in aircraft.