Ying Chen

Four-dimensional (4D) electronics, with the ability to deform over time in response to controlled actuation, have attracted rapidly growing interest for use in soft robots, biomedical devices, and wearable systems. A scheme to determine the amplitude and distribution of the actuation for the structure to deform into target shapes is essential for 4D electronics to operate precisely. Solution to this inverse problem is not straightforward in many cases, especially when nonlinearity governs the deformation. This work proposes a lightweight data-driven framework for the inverse problem in an electromagnetically actuated soft surface that takes the form of a Cu-EGaIn conductive network. The deformation of the soft surface is controlled by adjusting the applied voltage to regulate the current distribution and consequently the force distribution. Morphology measurement via the binocular vision method, combined with cubic interpolation, establishes the relationship between displacement of feature points and applied voltage. A machine learning architecture is then constructed with a balance between accuracy and training data volume, allowing for the inversion from surface morphology to predicted voltage. The framework avoids the construction of complex electromechanical coupling models and large deformation–current library databases. Using a relatively small experimental dataset of fewer than 300 samples, the prediction accuracies for the applied voltage to drive the morphable surface to target shapes are 82.91% for the activated interface position and 84.91% for the applied voltage value. Meanwhile, the proposed scheme was also preliminarily validated for cross-device reconstruction of liquid metal flexible films fabricated using the current manufacturing process. In addition, the data-driven scheme was further validated for inverse reconstruction beyond the training distribution through lightweight and rapid transfer learning. This work demonstrates the effectiveness of machine learning for determining excitation from surface morphology in electromagnetic actuation, serving as a foundation for 4D electronics.

Flexible integrated electronic technology enables the next generation of more flexible, freer, lighter, and smaller intelligent hardware, providing information sensing and edge computing capabilities for applications such as embodied intelligence, healthcare, and industrial Internet of Things. By thinning and integrating diverse chips onto a thin‐film polymer substrate, these flexible integrated devices retain advantages of semiconductor devices, including high integration density, low power consumption, and mature front‐end processes, without altering the fundamental device structures and materials. Although the electrical performance of ultra‐thin chips remains essentially unchanged after flexible packaging, the mechanical reliability during repeated bending deformation has been a critical concern due to the significant property mismatch between silicon and polymer materials, which may cause silicon fracture, interfacial failure, and eventual device failure. This study investigates and evaluates the fatigue lifetime of 20‐μm‐thick chips under repeated bending deformation with various radii of curvature. Through systematic experiments, the typical failure phenomena evolution is observed, finding three types of failure modes including metal pad wearing, metal stress migration, and gold wire break. The underlying failure mechanisms have been analyzed and explained by micro‐scale characterization. Finally, temperature effect on the fatigue lifetime has been explored trying to throw light on heating‐based accelerated fatigue life assessment method for the flexible chips. Hopefully, this work can provide essential support for the design to large‐scale application of flexible integrated devices in the widespread intelligent hardware products.

Touch links perception to control by turning contact into actionable force. It provides pressure and friction cues that help robots regulate grip and execute precise manipulation, especially in unstructured environments. As robots move into open, real‐world settings, rigid tactile sensors struggle with surface conformity. Flexible tactile sensors, with skin‐like softness and conformal contact, are emerging as a foundation for next‐generation robotic touch. This review provides a systematic survey of high‐impact advances in flexible tactile sensing for robotics, covering sensing mechanisms, structural design strategies, tactile decoding methods, and robotics‐oriented applications. We first analyze the characteristics, limitations, and robotic task suitability of different tactile sensing mechanisms, followed by a discussion of structural designs and integration approaches that enhance sensor performance. We then examine mechanics‐based reconstruction and intelligent decoding algorithms, including machine learning‐based tactile recognition and Transformer‐based tactile representation. Subsequently, we highlight frontier embodied intelligence applications enabled by flexible tactile sensing technologies. In addition, we identify four major research gaps and outline four priority directions for future development. This review aims to offer a robotics‐oriented, comprehensive technical reference for researchers in flexible tactile sensing and to promote substantive progress in robotic tactile perception.

Ultra-thin and flexible silicon chips are pivotal for high-performance conformal electronics. Backside grinding reduces chip thickness below 50 μm, enabling mass production of ultra-thin flexible chips in a low-cost way, yet the resulting defects and residual stress on the backside may propagate to the frontside circuits degrading the electric performance, especially for the critical-thickness devices. Here, commercial pulse-width modulation (PWM) chip wafers (involving tens of thousands of dies) based on bipolar transistors have been thinned to the thickness of ~ 20 μm by mechanical grinding with different parameters. Experimental results reveal a thickness-dependent bifurcated failure mechanism: Short-circuit current decays progressively with thickness reduction, while leakage current exhibits a catastrophic surge below the critical thickness (~ 18 μm). This work reveals a dual degradation mechanism in ultrathin ICs: Mechanical grinding not only amplifies substrate parasitic coupling via geometric thinning but also generates stress fields that induces dislocation rearrangement-aggregation cascades, ultimately dictating electrical failure modes. Chemical mechanical polishing (CMP) and reactive ion etching (RIE) have been deployed to inhibit leakage current surges by removing grinding-induced damaged layers and relieving interfacial residual stress, which collectively validate the stress-defect interaction as the governing mechanism of electrical failure in ultra-thin chips (UTCs). Hopefully, this study can throw light on the impact of mechanical grinding thinning on the electrical performance of ultra-thin chips paving the way to the wide applications of high-performance flexible electronics in the future.

Previous optogenetic bioelectronic systems have enabled a highly selective way of modulating neural populations by delivering a certain wavelength of light to engage with exogenously expressed light‐sensitive proteins, which lay the foundation of therapeutic interventions of neural circuits. However, real‐time biofeedback and strategic modulation are crucial for adjusting customized clinical treatment adjustment. To achieve this purpose, we integrated illumination, temperature, and electromyographic (EMG) sensing elements into the optogenetic bioelectronic system to avoid overexposure caused by localized overheating and to provide functional recovery evaluation during neural regeneration, which guides the in situ adjustment of the intensity, frequency, and duration of illumination parameters controlled by a wireless connected programmable external control board. In this study, both in vitro and in vivo experiments were performed to examine the optical, thermal, and electrical characteristics of our bioelectronic system. On this basis, we demonstrated a series of standardized EMG results to evaluate the recovery condition and modify the illumination parameters of each test rat. Combining temperature monitoring feedback and EMG signaling feedback, our optogenetic bioelectronic system enables strategic optogenetic spinal cord injury (SCI) treatment through real‐time illumination modulation to achieve customized spinal cord injury treatment.

Shape-changing smart substrates are revolutionizing implantable bioelectronics by offering minimal invasiveness and self-adaptive geometric conformity. However, precise manipulation of shapeshifting, which is critical for the successful implantation, are hindered by triggering accessibility and cumbersome customization. Here we introduce a digital 4D printing technique for fabricating customized shape memory hydrogel substrates enabling trigger-free shapeshifting with predefined onset periods. By controlling spatio-temporal light exposure, pixelated polymer sheets with heterogeneous networks are obtained, yielding intricate 3D original shapes after monomer removal. Meanwhile, spatially controlled onset time of the programmable shape recovery is introduced, attributed to local variation of polymeric network structure and the corresponding thermally-induced phase separation. This approach allows the complex 3D objects to execute autonomous multi-stage shape recovery through a single-step digital exposure process, which is not offered by existing techniques. In addition, the material exhibits negligible inflammation proven by in-vivo testing of male mice experiments. Demonstrated with self-winding bioelectronics, this method advances time-dependent adaptability of implantable devices and expands the functional scope of bioelectronics.

Owing to their excellent conformability and functional advantages derived from spatially induced structures, 3D conformal curvy electronics have garnered attention for their emerging applications in biomedical healthcare, soft machines, curvy imagers, etc. In this perspective, the historical evolution of 3D conformal curvy electronics is summarized, with representative examples highlighted and developmental trends outlined. The design strategies of 3D conformal curvy electronics are discussed across materials, structures, interfaces, and conformability assessment. Subsequently, diverse fabrication technologies are reviewed, including direct fabrication, conformal transfer printing, and conformal shape reconfiguration. Afterward, the typical applications of 3D conformal curvy electronics are presented, classified by integration with biological tissues, machines, and function-engineered curvy surfaces. Finally, the existing challenges and potential research directions are provided for further exploration.

Bionic water strider robots (BWSRs) have been a hot spot in research due to their unique biological inspiration and versatile applications in environmental monitoring. However, creating aquatic micro-robots with combined structure and function that can walk freely and perceive environmental information on water surfaces remains a challenge. Herein, self-propelled and integrated aquatic microrobots that can be remotely controlled by light and wireless environment detection are designed and fabricated by combining the actuating polydimethylsiloxane (PDMS)/carbon nanotube (CNT) substrate with heterogeneous wettability and the fully integrated wireless sensors. The substrate generates thrust on water through the photothermal Marangoni effect, serving as the propulsion system for this water-gliding microrobot. The heterogeneous wettability design features the robot with a hydrophobic body and four hydrophilic footpads. The hydrophobic body creates an air gap between the device and water, providing waterproof protection for electronics, making the robot move faster with smaller power and enabling greater load-carrying capacity. The hydrophilic footpads help the robot stand firmly on water, resisting overturn by waves and also enable modular assembly between robots. The load ability of a typical robot is 1,089% of its weight, resulting in the manufacture of microrobots with fully integrated visible light sensors and Bluetooth (BT) chips that can be steered by near-infrared laser. Hopefully, this strategy shall help to develop untethered aquatic robots with sophisticated actuation and wireless sensing for complex aquatic environments.

Aquatic actuators based on the light-to-work conversion are of paramount significance for the development of cutting-edge fields including robots, micromachines, and intelligent systems. Herein, we report the design and synthesis of near-infrared light-driven hydrogel actuators through loading with lightweight polydopamine-modified hollow glass microspheres (PDA-HGMPs) into responsive poly(N-isopropylacrylamide) (PNIPAM) hydrogels. These PDA-HGMPs can not only function as an excellent photothermal agent but also accelerate the swelling/desewlling of hydrogels due to their reconstruction for polymer gel skeleton, which speeds up the response rate of hydrogel actuators. The resulting hydrogel actuator shows controlled movements under light illumination, including complex self-propellant and floating/sinking motions. As the proof-of-concept demonstrations, a self-sensing robot is conceptualized by integrating the PDA-HGMP-containing hydrogel actuator with an ultrathin and miniature pressure sensor. Hopefully, this work can offer some important insights into the research of smart aquatic soft actuators, paving the way to the potential applications in emerging fields including micromachines and intelligent systems.