Search

AI edge devices require local intelligence for the concerns of latency and privacy. Given the accuracy and energy constraints, low-power convolutional neural networks (CNNs) are gaining popularity. To alleviate the high memory access energy and computational cost of large CNN models, prior works have proposed promising approaches including in-memory-computing (IMC) [1], mixed-signal multiply-and-accumulate (MAC) calculation [2], and reduced resolution network –[4]. With weights and activations restricted to ±1, binary neural network (BNN) combining with IMC greatly improves the storage and computation efficiency, making it well-suited for edge-based applications, and has demonstrated state-of-the-art energy efficiency in image classification problems [5]. However, compared to full resolution network, BNN requires larger model thus more operations (OPs) per inference for a certain accuracy. To address such challenge, we propose a mixed-signal ternary CNN based processor featuring higher energy efficiency than BNN. It confers several key improvements: 1) the proposed ternary network provides 1.5-b resolution (0/+1/-1), leading to 3.9x OPs/inference reduction than BNN for the same MNIST accuracy; 2) a 1.5b MAC is implemented by V CM -based capacitor switching scheme, which inherently benefits from the reduced signal swing on the capacitive DAC (CDAC); 3) the V CM -based MAC introduces sparsity during training, resulting in lower switching rate. With a complete neural network on chip, the proposed design realizes 97.1% MNIST accuracy with only 0.18uJ per classification, presenting the highest power efficiency for comparable MNIST accuracy.

Three-dimensional surface-conformable electronics is a burgeoning technology with potential applications in curved displays, bioelectronics, and biomimetics. Flexible electronics are notoriously difficult to fully conform to nondevelopable surfaces such as spheres. Although stretchable electronics can well conform to nondevelopable surfaces, they need to sacrifice pixel density for stretchability. Various empirical designs have been explored to improve the conformability of flexible electronics on spherical surfaces. However, no rational design guidelines exist. This study uses a combination of experimental, analytical, and numerical approaches to systematically investigate the conformability of both intact and partially cut circular sheets on spherical surfaces. Through the analysis of thin film buckling on curved surfaces, we identify a scaling law that predicts the conformability of flexible sheets on spherical surfaces. We also quantify the effects of radial slits on enhancing conformability and provide a practical guideline for using these slits to improve conformability from 40% to more than 90%.

E-skins consisting of soft pressure sensors are enabling technology for soft robots, bio-integrated devices, and deformable touch panels. A well-known bottleneck of capacitive pressure sensors (CPS) is the drastic decay in sensitivity with increasing pressure. To overcome this challenge, we have invented a hybrid-response pressure sensor (HRPS) that exhibits both the piezoresistive and piezocapacitive effects intrinsic to a highly porous nanocomposite (PNC) with carbon nanotube (CNT) dopants. The HRPS is constructed with two conductive electrodes sandwiching a laminated PNC and a stiff dielectric layer. We have simplified the hybrid response into a parallel resistor-capacitor circuit, whose output depends on the AC (alternating current) frequency used for the capacitance measurement. Herein, through theoretical analysis, we discover a dimensionless parameter that governs the frequency responses of the HRPS. The master curve is validated through experiments on the HRPS with various doping ratios, subject to different compressive strains, under diverse AC frequencies. In addition, the relative contribution of piezoresistive and piezocapacitive mechanisms are also found to vary with the three parameters. Based on this experimentally validated theory, we establish a very practical guideline for selecting the optimal AC frequency for the capacitance measurement of HRPSs.

Humans rely increasingly on sensors to address grand challenges and to improve quality of life in the era of digitalization and big data. For ubiquitous sensing, flexible sensors are developed to overcome the limitations of conventional rigid counterparts. Despite rapid advancement in bench-side research over the last decade, the market adoption of flexible sensors remains limited. To ease and to expedite their deployment, here, we identify bottlenecks hindering the maturation of flexible sensors and propose promising solutions. We first analyze challenges in achieving satisfactory sensing performance for real-world applications and then summarize issues in compatible sensor-biology interfaces, followed by brief discussions on powering and connecting sensor networks. Issues en route to commercialization and for sustainable growth of the sector are also analyzed, highlighting environmental concerns and emphasizing nontechnical issues such as business, regulatory, and ethical considerations. Additionally, we look at future intelligent flexible sensors. In proposing a comprehensive roadmap, we hope to steer research efforts towards common goals and to guide coordinated development strategies from disparate communities. Through such collaborative efforts, scientific breakthroughs can be made sooner and capitalized for the betterment of humanity. Agency for Science, Technology and Research (A*STAR) National Research Foundation (NRF) Submitted/Accepted version Y.L., Z.L., M.Z., and X.C. acknowledge the National Research Foundation, Singapore (NRF) under NRF’s Medium Sized Centre: Singapore Hybrid-Integrated Next-Generation μElectronics (SHINE) Centre funding programme, and AME programming funding scheme of Cyber Physiochemical Interface (CPI) project (no. A18A1b0045). Y.L. acknowledges National Natural Science Foundation of China (62201243). C.J. acknowledges funding support from the National Key R&D Program of China (no. 2019YFA0706100), the National Natural Science Foundation of China (82151305), Lingang Laboratory (LG-QS-202202-09). T.Q.T. and N.E.L. acknowledge support by the Basic Science Research Program (no. 2020R1A2C3013480) through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT. A.F. acknowledges the AFOSR (grant FA9550-22-1-0423). Y.L. and Y.Z. would like to acknowledge the NSF (award no. 2134664) and NIH (award no. R01HD108473) for financial support. X.F. acknowledges the support from the National Natural Science Foundation of China (grant no. U20A6001). L.Y. would like to thank the A*STAR Central Research Fund (CRF) and the AME Programmatic A18A1b0045 (Cyber Physiochemical Interfaces) for funding support. C.F.G. acknowledges the National Natural Science Foundation of China (no. T2225017). T.Q.T. acknowledges the Brain Pool Program (No. 2020H1D3A2A02111068) through the National Research Foundation (NRF) funded by the Ministry of Science. Z.L. acknowledges the support from RIE2020 AME Programmatic Grant funded by A*STAR-SERC, Singapore (Grant No. A18A1b0045). X.G. acknowledges funding support through the Shanghai Science and Technology Commission (grant no. 19JC1412400), the National Science Fund for Excellent Young Scholars (grant no. 61922057). C.D. acknowledges National Science Foundation CAREER: Conformable Piezoelectrics for Soft Tissue Imaging (grant no. 2044688) and MIT Media Lab Consortium funding. D.K. and O.G.S. acknowledge Leibniz Association and the German Research Foundation DFG (Gottfried Wilhelm Leibniz Program SCHM 1298/22-1, KA5051/1-1 and KA 5051/3-1), as well as the Leibniz association (Leibniz Transfer Program T62/2019). C.W. acknowledges the National Key Research and Development Program of China (grant no. 2021YFA1202600), National Natural Science Foundation of China (grant no. 62174082). A.V.-Y.T., E.Z., Y.Z., X.Z., and J.P. acknowledge the National Research Foundation, Singapore (NRF) under NRF’s Medium Sized Centre: Singapore Hybrid-Integrated Next-Generation μElectronics (SHINE) Centre funding programme, and AME programming funding scheme of Cyber Physiochemical Interface (CPI) project (no. A18A1b0045). R.Z. acknowledges National Natural Science Foundation of China (grant no. 51735007) and Beijing Natural Science Foundation (grant no. 3191001). N.M. acknowledges the support by JST PRESTO Grant Number JPMJPR20B7 and JST Adaptable and Seamless Technology transfer Program through Target-driven R&D (ASTEP) grant number JPMJTM22BK. C.P. acknowledges the Korean government (Ministry of Science and ICT, MSIT) (2022R1A4A3032923). M.W. acknowledges the National Key R&D Program of China under Grant (2021YFB3601200). X.Z. acknowledges National Natural Science Foundation of China (no. 62074029). S.X. acknowledges the 3M nontenured faculty award. T.-W.L. and D.-G.S. acknowledge the Pioneer Research Center Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT & Future Planning (grant no. NRF-2022M3C1A3081211). C.T.L. would like to acknowledge support from the Institute for Health Innovation and Technology (iHealthtech), the MechanoBioEngineering Laboratory at the Department of Biomedical Engineering and the Institute for Functional Intelligent Materials (I-FIM) at the National University of Singapore (NUS). C.T.L. also acknowledges support from the National Research Foundation and A*STAR, under its RIE2020 Industry Alignment Fund − Industry Collaboration Projects (IAF-ICP) (grant no. I2001E0059) − SIA-NUS Digital Aviation Corp Lab and the NUS ARTIC Research (grant no. HFM-RP1). X.Y. acknowledges funding support by City University of Hong Kong (grant no. 9667221). T.X. and X.Z. acknowledge National Natural Science Foundation of China (22234006). B.C.K.T. acknowledges Cyber-Physiochemical Interfaces CPI, A*STAR A18A1b0045. H.G. acknowledges a research start-up grant (002479-00001) from Nanyang Technological University and the Agency for Science, Technology and Research (A*STAR) in Singapore. W.G. acknowledges National Science Foundation grant 2145802. D.J.L. acknowledges support from the US National Science Foundation grant number CBET-2223566. G.Y. acknowledges support from The Welch Foundation award F-1861, and Camille Dreyfus Teacher-Scholar Award. M.D.D. acknowledges funding support from NSF (grant no. EEC1160483). J.-H.A acknowledges the National Research Foundation of Korea (NRF-2015R1A3A2066337). J.C. acknowledges the Henry Samueli School of Engineering & Applied Science and the Department of Bioengineering at the University of California, Los Angeles for startup support and a Brain & Behavior Research Foundation Young Investigator Grant. K.T. acknowledges JST AIP Accelerated Program (no. JPMJCR21U1) and JSPS KAKENHI (grant no. JP22H00594). P.S.W. acknowledges the National Science Foundation (CMMI1636136) for support. A.M.A., M.C.H., and P.S.W. thank the National Institute on Drug Abuse (DA045550) for support. S.M. and X.C. appreciated the support from the Smart Grippers for Soft Robotics (SGSR) Programme under the National Research Foundation, Prime Minister’s Office, Singapore under its Campus of Research Excellence and Technological Enterprise (CREATE) programme.

This paper introduces stretchable, long-term wearable, tattoo-like dry surface electrodes for highly repeatable electromyography (EMG). The tattoo-like sensors are hair thin, skin compliant and can be laminated on human skin just like a temporary transfer tattoo, which enables multi-day noninvasive but intimate contact with the skin even under severe skin deformation. The new electrodes were used to facilitate a system-based approach to tracking of long-term fatiguing and recovery processes in a human neuromusculoskeletal (NMS) system, which was based on establishing an autoregressive moving average model with exogenous inputs (ARMAX model) relating signatures extracted from the surface electromyogram (sEMG) signals collected using the tattoo-like sensors, and the corresponding hand grip force (HGF) serving as the model output. Performance degradation of the relevant NMS system was evaluated by tracking the evolution of the errors of the ARMAX model established using the data corresponding to the rested (fresh) state of any given subject. Results from several exercise sessions clearly showed repeated patterns of fatiguing and resting, with a notable point that these patterns could now be quantified via dynamic models relating the relevant muscle signatures and NMS outputs.

Adhesive ultrasound patches can provide medical imaging for patients on the go

When 2D materials are supported by substrates, matter trapped at the interface can coalesce to form nano- and microscale bubbles. These bubbles often negatively impact the performance of 2D material devices as they impede charge/photon/phonon transport across the interface. The difficulties created by these bubbles spurred research to understand how they form, whether their formation can be controlled, and what kind of matter is trapped inside them. These 2D material bubbles have since been exploited for novel chemistry and physics because of their ability to pressurize the trapped matter and strain the confining 2D material. The fabrication, characterization, and applications of 2D material bubbles are summarized in this review.

Soft pressure sensors are critical components of e-skins, which are playing an increasingly significant role in two burgeoning fields: soft robotics and bioelectronics. Capacitive pressure sensors (CPS) are popular given their mechanical flexibility, high sensitivity, and signal stability. After two decades of rapid development, e-skins based on soft CPS are able to achieve human-skin-like softness and sensitivity. However, there remain two major roadblocks in the way for practical application of soft CPS: the decay of sensitivity with increased pressure and the coupled response between in-plane stretch and out-of-plane pressure. In addition to existing strategies of building porous and/or high dielectric constant soft dielectrics, are there any other promising methods to overcome those bottlenecks? Are there any further considerations for the widespread deployment of e-skins? This perspective aims to shed some light on those topics.

Conventional electronic and photonic devices are inherently 2D and rigid because of the substrates on which they are fabricated. However, the world is not flat and stiff: There are many application...

Conventional wide-field-of-view cameras consist of multi-lens optics and flat image sensor arrays, which makes them bulky and heavy. As a result, they are poorly suited to advanced mobile applications such as drones and autonomous vehicles. In nature, the eyes of aquatic animals consist of a single spherical lens and a highly sensitive hemispherical retina, an approach that could be beneficial in the development of synthetic wide-field-of-view imaging systems. Here, we report an aquatic-vision-inspired camera that consists of a single monocentric lens and a hemispherical silicon nanorod photodiode array. The imaging system features a wide field of view, miniaturized design, low optical aberration, deep depth of field and simple visual accommodation. Furthermore, under vignetting, the photodiode array enables high-quality panoramic imaging due to the enhanced photodetection properties of the silicon nanorod photodiodes. By integrating a single monocentric lens with a hemispherical silicon nanorod photodiode array, a wide-field-of-view camera is created that offers low optical aberration, deep depth of field and simple visual accommodation.

This article presents a capacitively coupled voltage-controlled oscillator (VCO)-based sensor readout featuring a hybrid phase-locked loop (PLL)- $\Delta \Sigma $ modulator structure. It leverages phase-locking and phase-frequency detector (PFD) array to concurrently perform quantization and dynamic element matching (DEM), much-reducing hardware/power compared with the existing VCO-based readouts’ counting scheme. A low-cost in-cell data-weighted averaging (DWA) scheme is presented to enable a highly linear tri-level digital-to-analog converter (DAC). Fabricated in 40-nm CMOS, the prototype readout achieves 78-dB SNDR in 10-kHz bandwidth, consuming 4.68 $\mu \text{W}$ and 0.025-mm2 active area. With 172-dB Schreier figure of merit, its efficiency advances the state-of-the-art VCO-based readouts by $50\times $ .

In this work we consider bubbles that can form spontaneously when a two-dimensional (2D) crystal is transferred to a substrate with gases or liquids trapped at the crystal-substrate interface. The underlying mechanics may be described by a thin sheet on an adhesive substrate with the trapped fluid applying uniform transverse pressure. What makes this apparently simple problem complex is the rich interplay among geometry, interface, elasticity and instability. Particularly, extensive small-scale experiments have shown that the 2D crystal surrounding a bubble can adhere to and, meanwhile, slide on the substrate. The radially inward sliding causes hoop compression to the 2D crystal which may exploit wrinkling instabilities to relax or partially relax the compression. We present a theoretical model to understand the complex behaviors of even a linearly elastic 2D crystal due to the combination of nonlinear geometry, adhesion, sliding, and wrinkling in bubble systems. We show that this understanding not only successfully predicts the geometry of a spontaneous bubble but also reveals the strain-coupled physics of 2D crystals, e.g., the pseudomagnetic fields in graphene bubbles.

Bio-tissues are soft, curvilinear and dynamic whereas conventional electronics are hard, planar, and rigid. Our research focuses on the design, fabrication, conformability, and functionality of soft bio-integrated and bio-mimetic electronics based on inorganic functional materials such as metals, silicon, carbon nanotubes (CNT), and graphene. In particular, epidermal electronics, a.k.a. electronic tattoos (e-tattoos) represent a class of noninvasive stretchable circuits, sensors, and stimulators that are ultrathin, ultrasoft and skin-conformable. My group has invented a dry and freeform "cut-solder-paste" method for the rapid prototyping of multimodal, wireless, or very large area e-tattoos that are also high-performance and long-term wearable. The e-tattoos can be applied for physiological sensing, prosthesis control, and even human-mimetic robots.

High-performance wearable and implantable devices capable of recording physiological signals and delivering appropriate therapeutics in real time are playing a pivotal role in revolutionizing personalized healthcare. However, the mechanical and biochemical mismatches between rigid, inorganic devices and soft, organic human tissues cause significant trouble, including skin irritation, tissue damage, compromised signal-to-noise ratios, and limited service time. As a result, profuse research efforts have been devoted to overcoming these issues by using flexible and stretchable device designs and soft materials. Here, we summarize recent representative research and technological advances for soft bioelectronics, including conformable and stretchable device designs, various types of soft electronic materials, and surface coating and treatment methods. We also highlight applications of these strategies to emerging soft wearable and implantable devices. We conclude with some current limitations and offer future prospects of this booming field.

Implantation of biodegradable wafers near the brain surgery site to deliver anti-cancer agents which target residual tumor cells by bypassing the blood-brain barrier has been a promising method for brain tumor treatment. However, further improvement in the prognosis is still necessary. We herein present novel materials and device technologies for drug delivery to brain tumors, i.e., a flexible, sticky, and biodegradable drug-loaded patch integrated with wireless electronics for controlled intracranial drug delivery through mild-thermic actuation. The flexible and bifacially-designed sticky/hydrophobic device allows conformal adhesion on the brain surgery site and provides spatially-controlled and temporarily-extended drug delivery to brain tumors while minimizing unintended drug leakage to the cerebrospinal fluid. Biodegradation of the entire device minimizes potential neurological side-effects. Application of the device to the mouse model confirms tumor volume suppression and improved survival rate. Demonstration in a large animal model (canine model) exhibited its potential for human application., There is a need to further improve the efficacy of biodegradable wafers used in surgically treated brain tumors. Here, the authors report a flexible, biodegradable wireless device capable of adhesion to surgical site for optimal drug delivery upon mild-thermic actuation and report therapeutic efficacy in mouse and canine tumor models.

In van der Waals (vdW) heterostructures formed by stacking two monolayer semiconductors, lattice mismatch or rotational misalignment introduces an in-plane moire superlattice. While it is widely recognized that a moire superlattice can modulate the electronic band structure and lead to novel transport properties including unconventional superconductivity and insulating behavior driven by correlations, its influence on optical properties has not been investigated experimentally. We present spectroscopic evidence that interlayer excitons are confined by the moire potential in a high-quality MoSe2/WSe2 heterobilayer with small rotational twist. A series of interlayer exciton resonances with either positive or negative circularly polarized emission is observed in photoluminescence, consistent with multiple exciton states confined within the moire potential. The recombination dynamics and temperature dependence of these interlayer exciton resonances are consistent with this interpretation. These results demonstrate the feasibility of engineering artificial excitonic crystals using vdW heterostructures for nanophotonics and quantum information applications.

Unlike wafer-based rigid electronics, soft electronics have many unique advantages including thinness, flexibility, stretchability, conformability, lightweight, large area, as well as low cost. As a result, they have demonstrated many emerging capabilities in healthcare devices, soft robotics, and human-machine interface. Instead of conventional microfabrication, there is an ever-growing interest in the freeform or digital manufacture of soft electronics. This review provides a survey for a cost- and time-effective subtractive manufacturing process called the “cut-and-paste” method. It employs a mechanical cutter plotter to form patterns on various electronically functional membranes such as sheets of metals, functional polymers, and even two-dimensional (2D) materials, supported by a temporary tape. The patterned membranes can then be pasted on soft substrates such as medical tapes or even human skin. This process is completely dry and desktop. It does not involve any rigid wafers and is hence capable of making large-area electronics. The process can be repeated to integrate multiple materials on a single substrate. Integrated circuits (ICs) and rigid components can be added through a “cut-solder-paste” process. Multilayer devices can also be fabricated through lamination. We therefore advocate that the “cut-and-paste” method is a very versatile approach for the rapid prototyping of soft electronics for various applications.

We demonstrate sizable, unencapsulated e-tattoos with transformable serpentine design and Cartan curve–inspired transfer., Epidermal electrophysiology is widely carried out for disease diagnosis, performance monitoring, human-machine interaction, etc. Compared with thick, stiff, and irritating gel electrodes, emerging tattoo-like epidermal electrodes offer much better wearability and versatility. However, state-of-the-art tattoo-like electrodes are limited in size (e.g., centimeters) to perform electrophysiology at scale due to challenges including large-area fabrication, skin lamination, and electrical interference from long interconnects. Therefore, we report large-area, soft, breathable, substrate- and encapsulation-free electrodes designed into transformable filamentary serpentines that can be rapidly fabricated by cut-and-paste method. We propose a Cartan curve–inspired transfer process to minimize strain in the electrodes when laminated on nondevelopable skin surfaces. Unwanted signals picked up by the unencapsulated interconnects can be eliminated through a previously unexplored electrical compensation strategy. These tattoo-like electrodes can comfortably cover the whole chest, forearm, or neck for applications such as multichannel electrocardiography, sign language recognition, prosthetic control or mapping of neck activities.