|SUMMARY|||||AGENDA|||||SHORT COURSES|||||EXHIBITOR|||||CONFERENCE PRICING|||||NETWORKING OPPORTUNITIES|||||LOCATION|||||CFA|||||FLEXI AWARDS|
2018FLEX Student Posters
What are the bright minds of tomorrow developing in flexible and printed electronics? Be sure to stop by the Student Posters area outside the Exhibit Hall to learn more about the research coming out from these universities!
Sound Identification Using Physically-Expansive Sensing System
Levent E. Aygun, Ph.D. Candidate, Princeton University
Large-area electronics is a technology that provides a platform to build physically-expansive sensing systems with a flexible form factor. This enables the development of systems that can be seamlessly integrated into our everyday environment, facilitating collaborative spaces that improves the perception. One example is the spatially-distributed microphone array that enables gathering not only the sound signal information but also the sound source information. In this work, we show that the addition of the source location information can substantially improve the performance of machine learning algorithms used to identify the sound data. To this purpose, we created a physically-expansive sound sensing system (with spatially-distributed 15 microphones) that uses steered-response power using phase transform algorithm to predict the direction of the incoming sound signal. Predicted direction is combined with commonly used audio features to construct a feature vector used for the classification of the incoming sound. We have tested the performance of the sound identification using ESC-10 dataset and achieved considerable improvement with the addition of the incoming sound direction, obtaining 13% increase in the accuracy of support vector machine classifier and 34% increase in the accuracy of decision tree classifier. This work shows that sensor systems distributed over a physically-expansive space may ease the machine learning algorithms by perceiving the environment better than commonly used single sensor systems. We also addressed the computation and energy limitations for the realization of such systems and showed that audio feature extraction using genetic programming is promising for applications where energy consumption is critical.
Inkjet Printing of Dense Interconnect Arrays for Flexible Silicon Circuit Integration on Flexible Substrates
Jasmine Cox, Student Research Assistant, Boise State University
In this work, we demonstrate dense interconnect pattern printing for integrating and packaging Flexible Silicon circuits with printed circuits on flexible substrates. This silver interconnect pattern is intended to bridge the electrical connection between FleX^TM IC contact pads and electrode pads on the flexible substrate. Through optimizing the print parameters and curing process, 5 layers of silver lines are deposited to achieve dense 3.5 micron thick, 100 micron interconnect line arrays with a 50% duty cycle, and a feasible path towards increasing the density is presented. The thick lines also help in bridging the step height between the contacts on the flexible silicon circuit and those on the target PET substrate. The roll-to-roll compatible printed interconnect arrays will enable cost-efficient packaging of flexible hybrid electronic (FHE) systems.
Two-Phase Flow and Shell-Membrane Model for Imprint Lithography
Andrew Cochrane, Research Assistant, University of New Mexico
Jet-and-Flash imprint lithography is a high-throughput process being developed to replicate nano-featured patterns over large areas. It comprises four processing steps, the first being a drop-dispense process during which arrays of thousands of picoliter drops of photo-polymer are ink-jetted onto the substrate. The second step imprints these large drop arrays, resulting in drop merger and filling the nano-featured template while trapping and dissolving gas. Once the gas phase is eliminated, ultraviolet curing solidifies the liquid photopolymer in the shape of the pattern. Finally the pattern is separated from the cured polymer leaving behind a patterned surface and a thin residual layer. Minimizing gas-trapping and achieving residual layer thickness uniformity are two challenges to reliable pattern replication. A reduced-order model based on Reynolds' Lubrication theory is extended with the concept of relative permeability for coarse grain simulation of two-phase drop merger over large areas. The model incorporates the physics of gas compression and dissolution that give rise to the gas trapping problem. Simulation results are compared to a flow visualization to demonstrate its capability to relate degree of gas trapping to imprint gap thickness. Residual layer thickness non-uniformity is driven by conflicting capillary and lubrication forces and exacerbated by their increasing difference as the gap closes. The flow model is coupled with a shell membrane model to study how process parameters such as drop pattern, fluid properties and membrane stiffness affect the residual layer thickness that is formed by drops merging between a feature-less template and free-span tensioned web.
Fully Screen-Printed NiO Thermistor Arrays
Jonathan Ting, B.S. University of California, Berkeley
The performance of commercial Lithium ion batteries is a function of both battery voltage and temperature. Temperature has a significant effect on the performance and lifetime of these batteries, and influences several different failure mechanisms. One of the most dangerous failure mechanisms involves thermal runaway, where the battery becomes progressively more permanently damaged as it continues through each stage of thermal runaway. A possible method to improve battery safety is to monitor the temperature of the whole battery surface to potentially detect local heating, as a precursor to thermal runaway. Here, we present a fully screen-printed thermistor array, which uses a NiO based ink to sense temperature over a large area. We used the 4 x 4 array of thermistors to measure the temperature of a commercial Lithium ion battery when operated at higher C-rates. By multiplexing the individual thermistor pixels, we were able to plot, in real-time, the temperature of the battery surface, and spatially resolve local heating. Additionally, the screen-printing was done on flexible plastic substrates, making the sensor array conformal to the object being measured, making the sensor more efficient at resolving surface temperatures than conventional rigid thermistors.
Wireless, Flexible Bioelectronics-Enabled Brain-Machine Interface with Deep Convolutional Neural Networks
Musa Mahmood, Student, Georgia Institute of Technology
Due to natural variation in the human brain, electroencephalography (EEG) is difficult to implement into brain-machine interfaces (BMI). In addition, a conventional setup to record EEG suffers from low signal-to-noise ratio, while requiring an array of uncomfortable, rigid scalp electrodes with a massy hair cap. Recent advancements in classification of visually-evoked potentials via improved machine learning techniques have reduced the number of required electrodes from tens of channels to eight channels. However, the EEG measurement setup for multi electrodes still use a conductive gel and cap headgear to reduce the skin-contact impedance, which is a very obtrusive and cumbersome to use in real-life applications. Here, we introduce an ultralight, wireless, flexible bioelectronics on an elastomeric membrane that offers comfortable, long-term wearability on the skin. This device incorporates an ultrathin nanomembrane electrode on non-hair-bearing skin, flexible silver-composite carbon electrodes on hairy region, and miniaturized wireless electronics for data processing and transmission. Bluetooth-low-energy system and custom-designed Android application provides a real-time, portable monitoring of EEG, which can be easily integrated with the control of prosthetic devices and electronic wheelchairs. Furthermore, deep convolutional neural networks are used to map the occipital lobe region of the scalp using a conventional eight-channel EEG to find the optimal placement for only three-electrode configuration. Using this mobile, low-profile, and flexible EEG system with a combination of flexible- and skin-like electrodes, we can attain higher information transfer rates per data channel than other existing EEG systems. A Bluetooth-enabled system incorporating the flexible bioelectronics demonstrates a precise, hands-free control of a robotic wheelchair via EEG.
A Novel and Flexible Screen Printed Electrochemical Sensor on Paper Substrate
Dinesh Maddipatia, Ph.D. Student, Western Michigan University
A novel and flexible printed electrochemical sensor for the detection of various bio/chemicals was fabricated on a paper substrate using screen printing process. A two electrode sensor configuration with a working electrode (1700 µm radius) and a circular electrode (inner and outer radius of 2900 µm and 3900 µm) were printed on a wax-printed chromatography paper substrate using silver (Ag) flake ink. The performance of the printed electrochemical sensor was investigated by performing electrical impedance spectroscopy (EIS) to detect trace concentrations of glucose (C6H12O6). Impedance percentage changes of 10%, 24%, 34% and 49% for the 1 pM, 1 nM, 1 µM and 1 mM C6H12O6 solutions, respectively was obtained for the EIS based response when compared to deionized (DI) water at 20 Hz operating frequency and 1 mV applied potential. The responses of the printed electrochemical sensor demonstrated the feasibility of employing traditional screen printing process on a paper based substrate for bio/chemical sensing applications.
Development of a flexible strain gauge on paper based platform using additive printing process
Dinesh Maddipatia, Ph.D. Student, Western Michigan University
A novel printed strain gauge, for structural monitoring applications, was fabricated on a flexible paper substrate using additive flexography printing process. A meander trace with length and width of 126 mm and 0.4 mm, respectively was printed on the paper substrate using silver (Ag) flexo ink. The capability of the printed strain gauge was investigated by subjecting it to a cyclic 3- point bend test, with a displacement of 1 mm and 2 mm at 3 Hz operating frequency for 500 cycles. An overall resistance change of 6.4 % and 6.5 % was observed in the base resistance and bend resistance, respectively for the electro-mechanical response of the strain gauge towards 1 mm displacement after 500 cycles of bending. Similarly, an 87.97 % and 28.8 % resistance change was observed for the base resistance and bend resistance, respectively after 500 cycles of bending for 2 mm displacement. The responses of the paper based strain gauge demonstrated the feasibility of employing additive flexography print manufacturing process for the development of an efficient, flexible and cost-effective device for structural monitoring applications.
A Fully Printed Carbon Nanotube (CNT) Based Thermistor
Vikram Shreeshail Turkani, Student, Western Michigan University
A fully printed carbon nanotube (CNT) based thermistor has been successfully fabricated on flexible polyethylene terephthalate (PET) substrate using screen and gravure printing processes. The thermistor consists of screen printed bottom silver electrodes, a gravure printed active layer (CNT) on top of the bottom silver electrode, a screen printed organic encapsulation layer on top of the active layer, and a secondary screen printed silver based encapsulation layer on top of the organic encapsulation layer. The capability of the fabricated thermistor was investigated by measuring its response towards varying temperatures and relative humidity (RH). The resistive response of the thermistor demonstrated a linear relationship for temperatures varying from 30 °C to 100 °C. An average sensitivity of 2 kΩ/°C, with a correlation coefficient of 0.99, was obtained for the printed thermistor. In addition, the resistive response of the thermistor towards humidity ranging from 20 % RH to 70 % RH, in steps of 10 % RH for two temperatures 30 ⁰C and 50 ⁰C, was investigated. A maximum resistance change of 0.1 % and 0.34 % was observed at 40 % RH and 70 % RH, for temperatures of 30 ⁰C and 50 ⁰C, respectively when compared to the base resistance at 20 % RH.
Mm-wave Ultra-Long-Range Energy-Autonomous Printed RFID Van-Atta Wireless Gas Sensors: at the Crossroads of 5G and IoT
Jimmy Hester, Ph.D. Candidate, Georgia Institute of Technology
The rapid, sustained, and concurrent growths of the Internet of Things (IoT) and 5G communications networks are two of the most significant modern trends in wireless technologies. Nevertheless, hardly any avenues for overlap between both fields have yet been proposed. This can be explained by the apparent contradiction between the requirement for costly and power-hungry mm-wave technologies (5G) on the one hand, and the relentless search for low cost solutions (IoT) on the other. Here, an architecture that encompasses the best of both worlds is reported, in the shape of the first printed mm-wave RFID sensor. By leveraging the respective advantages of both technologies, this inkjet-printed flexible mm-wave smart skin provides an unprecedented solution for the emergence of ultra-long-range RFID sensors, interfaced by soon-to-be-pervasive 5G networks. In order to achieve this, a mm-wave RFID tag is constructed on a flexible LCP substrate, around a fully-inkjet-printed Van-Atta reflectarray structure. In addition, low-cost single-transistor switches and a timer were added to enable a backscatter communications scheme, with the modulation frequency related to the state of a fully-inkjet-printed resistometric carbon-nanotubes-based ammonia sensor. The entire system, displaying a power consumption of under 0.3 mW, is battery-less and powered by a flexible solar cell. The reported prototype, as the first example of a mm-wave energy-autonomous low-cost gas sensing node, may enable the emergence of fully-printed ubiquitous and 5G-interfaced smart skins of the IoT.
All-Soft and 3D-Integrated Functional Microsystems Enabled by Gallium-Based Liquid Metal and Soft Lithography
Min-gu Kim, Ph.D. Candidate, Georgia Institute of Technology
Lightweight, flexible, and stretchable wearable electronics have gained significant attention for various sensing applications ranging from entertainment to healthcare, but the mechanical mismatch between soft biological skins and conventional rigid and bulky electronic materials often limits the ultimate usability and leads to hard-soft material interface failure. To circumvent this limitation, the use of conducting liquid, such as gallium-based liquid metal (eutectic gallium-indium alloy, EGaIn), has great potential because of its non-toxicity, low melting temperature, and excellent electrical and mechanical properties. However, EGaIn patterning challenges, particularly regarding minimum feature sizes, size-scalability, uniformity, and residue-free surfaces, have limited the demonstration of high-density, soft microelectronic devices. This research investigates an advanced EGaIn thin-film patterning based on soft lithography and a compatible vertical integration technique, which enable size-scalable and high-density EGaIn-based soft microelectronics. The proposed patterning technique overcomes the current limitation in EGaIn fabrication by demonstrating uniform and residue-free EGaIn thin lines with width from single micrometers to several millimeters at room temperature and under ambient pressure. Also, vertical integration using EGaIn-filled soft vias facilitates high-density integration as well as system-level flexibility and stretchability. To this end, by combining the scalable fabrication process and a vertical integration approach, two types of all-soft and 3D-integrated microsystems are demonstrated: i) a finger-mountable strain sensing microsystem with reduced temperature sensitivity and ii) wireless and battery-free chemical microsystems for liquid- and gas-phase volatile organic compound (VOC) sensing. The demonstrated fabrication and integration approaches provide a path towards all-soft and highly integrated wearable physical and chemical microsystems for human-machine interfaces, personal environmental and healthcare monitoring applications.
Recent Advances of RF Energy Harvesting for IoT Devices usign 3D/Inkjet Printed Technology
Aline Eid, Ph.D. Student, Georgia Institute of Technology
In this work, the particular importance and associated opportunities of additively manufactured radio frequency (RF) components and RF energy harvesters for IoT devices are discussed. This effort addresses some of the key bottlenecks for the realization of efficient harvesting systems. RF energy harvesting networks have been drawing significant attention because of their potential to power millions of autonomous devices using ambient RF energy. It has been proposed as the main energy source for low power consumption and low duty cycle devices. Some of the challenges of ambient RF energy harvesting are the requirement of operation over a multitude of frequency bands of low ambient power densities resulting in a very wide aggregate operating bandwidth, the operation at very low input power levels resulting with low efficiencies, the allocation of the best harvesting sources available and the accurate prediction of the harvester’s position. This work demonstrates the possibility of the integration of additive manufacturing technologies (AMTs) such as inkjet/3D‐printing for the ultra-low cost, on-demand reconfigurable and scalable fabrication of flexible wearable RF energy harvesters. It covers the use of inkjet-printed wearable and flexible energy harvesters along with the integration of machine learning techniques for energy autonomous on-body sensing networks, fully autonomous ultra-low power hybrid RF/photovoltaic energy harvesting system and ultralightweight multiband RF energy harvesters. This system implementation allows for much enhanced harvesting capabilities and helps to overcome several challenges of ambient RF energy harvesting.
Sensorized Pneumatic Muscles for Force and Stiffness Control
Lucas Tiziani, Graduate Student, Georgia Institute of Technology
This poster presents the design and experimental validation of a soft pneumatic artificial muscle with position and force sensing capabilities. Conductive liquid-based soft sensors are embedded in a fiber-reinforced contractile actuator to measure two modes of deformation – axial strain and diametral expansion – which, together, are used to determine the stroke length and contractile force generated under internal pressure. The proposed device was validated by using data from the embedded sensors to estimate the force output of the actuator at fixed lengths, as well as the stiffness and force output of a one degree-of-freedom hinge joint driven by an antagonist pair of the sensorized pneumatic muscles.
Analysis of Microplasm Discharge on a Flexible Platform for Sterilization Applications
Arnesh Bose, Student, Western Michigan University
In this study, a microplasma discharge based device was modeled and simulated in Argon (Ar) based environment with varying ambient conditions for sterilization applications. The device was modeled with copper and dielectric based electrodes and flexible thin film polyethylene terephthalate (PET) based substrate. The model of the device was simulated in COMSOL Multiphysics using plasma module, to investigate the effective microplasma discharge at an applied potential of 2031 V and frequency of 17.26 Hz. Initially, the ambient pressure was kept constant at 1 atm and the temperature was varied from 240K to 360K. Electron densities and mobilities ranging from 1.35 × 1013 m-3 to 1.57 × 1013 m-3 and 0.13 [m2/(V.s)] to 0.19 [m2/(V.s)] was observed as the temperature was increased and then decreased from 240K to 360K, respectively. Then, a simulation was performed with atmospheric pressure conditions ranging from 0.3 atm to 1.3 atm and a constant temperature of 295.15K (22°C). As the pressure was increased and then decreased from 0.3 atm to 1.3 atm, electron densities and mobilities ranging from 1.35 × 1013 m-3 to 1.57 × 1013 m-3 and 0.13 [m2/(V.s)]to 0.19 [m2/(V.s)], respectively was obtained. These results demonstrate the capability of the model to generate microplasma discharge at varying ambient temperature and pressure.
Soft Fluidic Modulation of Skin Temperature
Donald Ward, Graduate Research Assistant, Georgia Institute of Technology
A non-invasive, soft, thermal vaso-modulator system was designed and evaluated, for use in physiological experiments to assess neuro-vascular health. The system stimulates biomechanisms involved in thermoregulation by varying the temperature of the skin with a soft, fluidically-controlled thermal pad. The silicone-based thermal pad was strategically embedded with graphite particles to improve the base thermal conductivity. The temperature control mechanism was achieved by implementing a proportional-integral control loop with feedback from the heating pad touching the skin. To verify the system operation, physiological signals were collected from a human subject to monitor local blood volume changes in response to temperature modulation. The mechanisms associated with thermoregulation were quantified by extracting features from the measured waveforms.