김동하 교수

Abstract Physical neural networks made of analog resistive switching processors are promising platforms for analog computing. State-of-the-art resistive switches rely on either conductive filament formation or phase change. These processes suffer from poor reproducibility or high energy consumption, respectively. Herein, we demonstrate the behavior of an alternative synapse design that relies on a deterministic charge-controlled mechanism, modulated electrochemically in solid-state. The device o

Segregation of aliovalent dopant cations is a common degradation pathway on perovskite oxide surfaces in energy conversion and catalysis applications. Here we focus on resolving quantitatively how dopant segregation is affected by oxygen chemical potential, which varies over a wide range in electrochemical and thermochemical energy conversion reactions. We employ electrochemical polarization to tune the oxygen chemical potential over many orders of magnitude. Altering the effective oxygen chemic

The SnO(2) anode is a promising anode for next-generation Li ion batteries because of its high theoretical capacity. However, it exhibits inherent capacity fading because of the large volume change and pulverization that occur during the charge/discharge cycles. The buffer matrix, such as electrospun carbon nanofibers (CNFs), can alleviate this problem to some extent, but SnO(2) particles are thermodynamically incompatible with the carbon matrix such that large Sn agglomerates form after carboni

Abstract The instability of the surface chemistry in transition metal oxide perovskites is the main factor hindering the long-term durability of oxygen electrodes in solid oxide electrochemical cells. The instability of surface chemistry is mainly due to the segregation of A-site dopants from the lattice to the surface. Here we report that cathodic potential can remarkably improve the stability in oxygen reduction reaction and electrochemical activity, by decomposing the near-surface region of t

The electrochemical reduction of CO 2 in acidic media offers the advantage of high carbon utilization, but achieving high selectivity to C 2+ products at a low overpotential remains a challenge. We identified the chemical instability of oxide-derived Cu catalysts as a reason that advances in neutral/alkaline electrolysis do not translate to acidic conditions. In acid, Cu ions leach from Cu oxides, leading to the deactivation of the C 2+ -active sites of Cu nanoparticles. This prompted us to desi

Controlling the size of Au nanoparticles (NPs) and their interaction with the oxide support is important for their catalytic performance in chemical reactions, such as CO oxidation and water-gas shift. It is known that the oxygen vacancies at the surface of support oxides form strong chemical bonding with the Au NPs and inhibit their coarsening and deactivation. The resulting Au/oxygen vacancy interface also acts as an active site for oxidation reactions. Hence, small Au NPs are needed to increa

Applying anodic potential can be an efficient way to re-activate the perovskite oxide surface by incorporating the surface dopant precipitates into the perovskite phase.

Abstract Gas sponges capable of absorbing, storing, and releasing ions in a reversible manner are in high demand for advanced electronics, energy devices, and sensors. Here, it is shown that brownmillerite BaInO 2.5 epitaxial films exhibit the capability to act as solid‐state catalytic hydrogen sponges at a remarkably low temperature (≈100 °C). Compared to sintered pellets with random crystallographic orientations and many defects, BaInO 2.5 epitaxial films give three orders of magnitude higher

The Internet of things (IoT) integrates heterogeneous computing devices, allowing each node to communicate with one another. However, the connected “things” raise security challenges that need protection for IoT devices from network-based attacks. As an integrated solution, Secure Swarm Toolkit (SST) provides authorization infrastructure that addresses the security requirements of IoT devices. The pre-release version of SST primarily provided the Node.js and JavaScript-based API for programming

Owing to its pseudocapacitive, unidimensional, rapid ion channels, TiO 2 (B) is a promising material for application to battery electrodes. In this study, we align these channels by epitaxially growing TiO 2 (B) films with the assistance of an isostructural VO 2 (B) template layer. In a liquid electrolyte, binder-free TiO 2 (B) epitaxial electrodes exhibit a supercapacity near the theoretical value of 335 mA h g –1 and an excellent charge–discharge reproducibility for ≥200 cycles, which outperfo

Abstract The interest in highly sensitive sensors is rapidly increasing for detecting very tiny signals for Internet of Things devices. Here, we achieve ultra-sensitive correlated breathable sensors based on freestanding VO 2 membranes. We fabricate the membranes by growing VO 2 films onto sacrificial Sr 3 Al 2 O 6 layer grown on SrTiO 3 , selectively dissolving the Sr 3 Al 2 O 6 in water, and then rendering freestanding VO 2 membrane on nanomesh. The nanomeshes are extremely flexible, sweat per

This article studies the role of architecture design, i.e. choice of the number of nodes at each hidden layer, in deep neural networks (DNNs). We give a theoretical explanation that invariance and complexity of a DNN are determined by the design of its architecture. To be more specific, for DNNs with the rectified linear activation function, we prove that the variations of gradients become the largest when the bottleneck layer, the layer with the fewest nodes, changes its activation pattern and