Searcharxiv⌕ Search

arXiv subjects

I-Wei Chen

Publications and source records attributed to I-Wei Chen.

At least 19 recordsLinked to original sources
arXiv2025

Team NYCU at Defactify4: Robust Detection and Source Identification of AI-Generated Images Using CNN and CLIP-Based Models

With the rapid advancement of generative AI, AI-generated images have become increasingly realistic, raising concerns about creativity, misinformation, and content authenticity. Detecting such images and identifying their source models has become a critical challenge in ensuring the integrity of digital media. This paper tackles the detection of AI-generated images and identifying their source models using CNN and CLIP-ViT classifiers. For the CNN-based classifier, we leverage EfficientNet-B0 as the backbone and feed with RGB channels, frequency features, and reconstruction errors, while for CLIP-ViT, we adopt a pretrained CLIP image encoder to extract image features and SVM to perform classification. Evaluated on the Defactify 4 dataset, our methods demonstrate strong performance in both tasks, with CLIP-ViT showing superior robustness to image perturbations. Compared to baselines like AEROBLADE and OCC-CLIP, our approach achieves competitive results. Notably, our method ranked Top-3 overall in the Defactify 4 competition, highlighting its effectiveness and generalizability. All of our implementations can be found in https://github.com/uuugaga/Defactify_4

cs.CV↗
arXiv2020

DC Resistance Degradation of SrTiO$_3$: The Role of Virtual-Cathode Needles and Oxygen Bubbles

This study of highly accelerated lifetime tests of SrTiO$_3$, a model semiconducting oxide, is motivated by the interest in reliable multilayer ceramic capacitors and resistance-switching thin-film devices. Our analytical solution to oxygen-vacancy migration under a DC voltage -- the cause of resistance degradation in SrTiO$_3$ -- agrees with previous numerical solutions. However, all solutions fail to explain why degradation kinetics feature a very strong voltage dependence, which we attribute to the nucleation and growth of cathode-initiated fast-conducting needles. While they have no color contrast in SrTiO$_3$ single crystals and are nominally invisible, needles presence in DC-degraded samples -- in silicone oil and in air -- was unambiguously revealed by in-situ hot-stage photography. Observations in silicone oil and thermodynamic considerations of voltage boundary conditions further revealed a cooccurrence of copious oxygen bubbling and the onset of final accelerating degradation, suggesting sudden oxygen loss is a precursor of final failure. Remarkably, both undoped and Fe-doped SrTiO$_3$ can emit electroluminescence at higher current densities, thus providing a vivid indicator of resistance degradation and a metal-to-insulator resistance transition during cooling. The implications of these findings to thin ceramic and thin film SrTiO$_3$ devices are discussed, along with connections to similar findings in likewise degraded fast-ion yttria-stabilized zirconia.

cond-mat.mtrl-sci↗
arXiv2020

Ultra-uniform Nanocrystalline Materials via Two-Step Sintering

Nanocrystalline metals and ceramics with <100 nm grain sizes and superior properties (e.g., mechanical strength, hardness, fracture toughness and stored dielectric energy) are of great interest. Much has been discussed about achieving nano grains, but little is known about maintaining grain-size uniformity that is critical for material reliability. An especially intriguing question is whether it is possible to achieve a size distribution narrower than what Hillert[1] theoretically predicted for normal grain growth, a possibility suggested, for growth with a higher growth exponent, by the generalized mean-field theory[2] of Lifshitz, Slyozov, Wagner (LSW)[3,4] and Hillert but never realized in practice. We demonstrate that this can be achieved in bulk materials with an appropriately designed two-step sintering route that (a) takes advantage of the large growth exponent in the intermediate sintering stage to form a most uniform microstructure despite porosity remaining, and (b) freezes the grain growth thereon while continuing densification to reach full density. The resultant dense bulk Al2O3 ceramic has an average grain size of 34 nm and a much narrower size distribution than Hillert's prediction. Bulk Al2O3 with a grain-size distribution narrower than the particle-size distribution of starting powders was also demonstrated using this strategy, as were highly uniform bulk engineering metals and ceramics of either high purity and high melting points (Mo and W-Re) or highly complex compositions (core-shell BaTiO3 and 0.87BaTiO3-0.13Bi(Zn2/3(Nb0.85Ta0.15)1/3)O3).

cond-mat.mtrl-sci↗
arXiv2020

Potential jumps at transport bottlenecks cause instability of nominally ionic solid electrolytes in electrochemical cells

Normal operations of electrochemical devices such as solid oxide fuel cells (SOFC), solid oxide electrolyzer cells (SOEC) and lithium ion batteries (LIB) sometimes fail because of unexpected formation of internal phases. These phases include oxygen bubbles at grain boundaries inside the zirconia electrolyte of SOEC, isolated Li metal islands inside the (garnet type) Li7La3Zr2O12 electrolyte of all-solid-state LIB, and similar Na metal islands inside the Na-beta-alumina and NASICON electrolytes of Na-S batteries. Remarkably, although the devices can operate in both polarities, the propensity for failure depends on the polarity. Here we explain these and other phenomena in nominally ionic solid electrolytes and mixed-conducting electrodes in simple thermodynamic and kinetic terms: the unexpected internal phases are caused by a large potential jump that is needed to push a constant ion or electron flow through its internal transport bottleneck. Definite rules for internal phase formation including its polarity dependence are formulated to help predict and mitigate it, which leads to microstructural instability, efficiency deterioration and breakdown.

cond-mat.mtrl-sci↗
arXiv2019

Electron Localization Enhances Cation Diffusion in Transition Metal Oxides: An Electronic Trebuchet Effect

Ion diffusion is a central part of materials physics of fabrication, deformation, phase transformation, structure stability and electrochemical devices. Conventional theory focuses on the defects that mediate diffusion and explains how their populations influenced by oxidation, reduction, irradiation and doping can enhance diffusion. However, we have found the same influences can also elevate their mobility by orders of magnitude in several prototypical transition-metal oxides. First-principles calculation fundamentally connects the latter observation to migrating ion's local structure, which is inherently soft and has a broken symmetry, making it susceptible to electron or hole localization, thereby realizing a lower saddle-point energy. This finding resolves an unanswered question in physical ceramics of the past 30 years: why cation diffusion against the prediction of classical nonstoichiometric defect physics is enhanced in reduced zirconia, ceria and structurally related ceramics? It also suggests the saddle-point electron-phonon interaction that enables a negative-U state is akin to the counterweight effect that enables a trebuchet. This simple picture for the transitional state explains why enhanced kinetics mediated by radical-like-ion migration occurs often, especially under extreme conditions.

cond-mat.mtrl-sci↗
arXiv2019

DC Electrical Degradation of YSZ: Voltage Controlled Electrical Metallization of A Fast Ion Conducting Insulator

DC electrical degradation as a form of dielectric and resistance breakdown is a common phenomenon in thin-film devices including resistance-switching memory. To obtain design data and to probe the degradation mechanism, highly accelerated lifetime tests (HALT) are often conducted at higher temperatures with thicker samples. While the mechanism is well established in semiconducting oxides such as perovskite titanates, it is not in stabilized zirconia and other fast oxygen-ion conductors that have little electronic conductivity. Here we model the mechanism by an oxygen-driven, transport-limited, metal-insulator transition, which finds support in rich experimental observations - including in situ videos and variable temperature studies - of yttria-stabilized zirconia. They are contrasted with the findings in semiconducting titanates and resistance memory, and provide new insight into ceramic processing with extremely rapid heating and cooling such as flash sintering and melt processing.

cond-mat.mtrl-sci↗
arXiv2018

Inversion of oxygen potential transitions at grain boundaries of SOFC/SOEC electrolytes

Solid oxide fuel/electrolyzer cell (SOFC/SOEC) converts energy between chemical and electrical forms inversely. Yet electrolyte degradation takes place much more severe for SOEC than SOFC during long-term operations. By solving transport equations, we found very large oxygen potential gradients and sharp oxygen potential transitions at grain boundaries of polycrystalline SOFC/SOEC electrolyte. Surprisingly, an inversion of oxygen potential transitions was identified, suggesting a fundamentally different transport mechanism for minor electronic charge carriers. Such findings could be critical to understand and eliminate SOFC/SOEC degradations in practical applications.

cond-mat.mtrl-sci↗
arXiv2018

Electron Localization Enhances Cation Diffusion in Reduced ZrO2, CeO2 and BaTiO3

According to defect chemistry, the experimental observations of enhanced cation diffusion in a reducing atmosphere in zirconia, ceria and barium titanate are in support of an interstitial mechanism. Yet previous computational studies always found a much higher formation energy for cation interstitials than for cation vacancies, which would rule out the interstitial mechanism. The conundrum has been resolved via first-principles calculations comparing migration of reduced cations and oxidized ones, in cubic ZrO2, CeO2 and BaTiO3. In nearly all cases, reduction alone lowers the migration barrier, and pronounced lowering results if cation's electrostatic energy at the saddle point decreases. The latter is most effectively realized when a Ti cation is allowed to migrate via an empty Ba site thus being fully screened all the way by neighboring anions. Since reduction creates oxygen vacancies as well, which are highly mobile, we also studied their effect on cation migration, and found it only marginally lowers the migration barrier. In several cases, however, a large synergistic effect between cation reduction and oxygen vacancy is revealed, causing an electron to localize in the saddle-point state at a much lower energy than normal, signaling that the saddle point is a negative-U state in which the soft environment enables a large electron-phonon interaction that can over-compensate the on-site Coulomb repulsion. These general findings are expected to be applicable to defect-mediated ion migration in most transitional metal oxides.

cond-mat.mtrl-sci↗
arXiv2018

Oxygen Potential Transition in Mixed Conducting Oxide Electrolyte

It is generally assumed that oxygen potential in a thin oxide electrolyte follows a linear distribution between electrodes. Jacobsen and Mogensen have shown, however, that this is not the case for thin zirconia membranes in solid oxide electrochemical cells. Here we demonstrate that there is a ubiquitous oxygen potential transition rooted in the p-type/n-type transition of electronic conductivity inside mixed conducting oxides, and that the transition is extremely sensitive to electrode potential and current density. It is also remarkably sensitive to the conductivity ratio of electrons and holes, as well as their association with lattice oxygens and vacancies, which tends to increase the oxygen flow. Direct evidence of a sharp oxygen potential transition has been found in an equally sharp grain size transition in electrically loaded zirconia. More broadly speaking, the oxygen potential transition is akin to a first-order phase transition. Therefore, it will suffer interface instability, especially in high-current-density devices. These findings provide new opportunities to understand several disparate observations in the literature, from microstructural degradation and stress distribution in solid oxide fuel/electrolyzer cells, to field-assisted sintering, to conducting filaments in resistance memory, to dendrite formation in electrochemical cells.

cond-mat.mtrl-sci↗
arXiv2018

A Si-memristor electronically and uniformly switched by a constant voltage

Amorphous insulators have localized wave functions that decay with the distance $r$ following exp($-r/ζ$). Since nanoscale conduction is not excluded at $r<ζ$, one may use amorphous insulators and take advantage of their size effect for nanoelectronic applications. Voltage-regulated nanoscale conductivity is already utilized in metal-insulator-metal devices known as memristors. But typically their tunable conductivity does not come from electrons but from migrating ions within a stoichastically formed filament, and as such their combined resistor-memory performance suffers. Here we demonstrate amorphous-silicon-based memristors can have coherent electron wave functions extending to the full device thickness, exceeding 15 nm. Remarkably, despite the large aspect ratio and very thin thickness of the device, its electrons still follow an isotropic, three-dimensional pathway, thus providing uniform conductivity at the nanometer scale. Such pathways in amorphous insulators are derived from overlapping gap states and regulated by trapped charge, which is stabilized by electron-lattice interaction; this makes the memristor exhibit pressure-triggered insulator$\rightarrow$metal transitions. Fast, uniform, durable, low-power and purely electronic memristors with none of the shortcomings of ion-migrating memristors have been fabricated from a variety of amorphous silicon compositions and can be readily integrated into silicon technology. Therefore, amorphous silicon may provide the ideal platform for building proximal memories, transistors and beyond.

cond-mat.mes-hall↗
arXiv2018

Purely electronic nanometallic ReRAM

Resistance switching random access memory (ReRAM), with the ability to repeatedly modulate electrical resistance, has been highlighted as a feasible high-density memory with the potential to replace negative-AND (NAND) flash memory. Such resistance modulation usually involves ion migration and filament formation, which usually lead to relatively low device reliability and yield. Resistance switching can also come from an entirely electronic origin, as in nanometallic memory, by electron trapping and detrapping. Recent research has revealed additional merits of its mechanism, which entails smart, atomic-sized floating gates that can be easily engineered in amorphous Si, oxides, and nitrides. This article addresses the basic ideas of nanometallic ReRAM, which may also be a contender for analogue computing and non-von Neumann-type computation.

cond-mat.mes-hall↗
arXiv2017

Electrical and Hydrogen Reduction Enhances Kinetics in Doped Zirconia and Ceria: II. Mapping Electrode Polarization and Vacancy Condensation in YSZ

Knowing the correlation between grain boundary mobility and oxygen potential in yttria stabilized zirconia (YSZ), we have utilized the grain size as a microstructural marker to map local oxygen potential. Abrupt oxygen potential transition is established under a large current density and in thicker samples. Cathodically depressed oxygen potential can be easily triggered by poor electrode kinetics or in an oxygen-lean environment. Widespread cavitation in the presence of highly reducing oxygen potential suggests oxygen vacancy condensation instead of oxygen bubble formation as commonly assumed for solid oxide fuel/electrolysis cells. These results also suggest electrode kinetics has a direct influence on the microstructure and properties of ceramics sintered under a large electric current.

cond-mat.mtrl-sci↗
arXiv2017

Grain Growth with Size-Dependent or Statistically Distributed Mobility

Conventional grain growth is rate-limited by the mobility of grain boundary. To describe similar phenomena limited by the mobility of other grain junctions, we have developed a general theory allowing for size-dependent mobility and its statistical variance. We obtained analytic solutions for the steady-state size distribution and the growth exponent, defined as (grain size)n ~ time, down to n=1, which arises when the mobility of three-grain lines is rate-limiting. When the mobility of four-grain junctions is rate-limiting, exponential growth and a bifurcating size distribution result. These solutions manifest a general trend: The size distribution narrows with increasing n. Yet experimentally the opposite trend has been observed recently, which can only be reproduced in simulation if the mobility distribution is made at lease bimodal, with one mode being immobile or nearly immobile. The latter can be realized in slow grain growth below the temperature of mobility transition.

cond-mat.mtrl-sci↗
arXiv2017

Mobility Transition at Grain Boundaries in Two-Step Sintered 8 mol% Yttria Stabilized Zirconia

Stagnation of grain growth is often attributed to impurity segregation. Yttria-stabilized cubic zirconia does not evidence any segregation-induced slowdown, as its grain growth obeys the parabolic law when the grain size increases by more than one order of magnitude. However, lowering the temperature below 1300 oC triggers an abrupt slowdown, constraining the average grains to grow by less than 0.5 $μ$m in 1000 h despite a relatively large driving force imparted in the fine grains of ~0.5 $μ$m. Yet isolated pockets of abnormally large grains, along with pockets of abnormally small grains, emerge in the same latter sample. Such microstructure bifurcation has never been observed before, and can only be explained by an inhomogeneous distribution of immobile four-grain junctions. The implications of these findings for two-step sintering are discussed.

cond-mat.mtrl-sci↗
arXiv2017

A Computational Study of Yttria-Stabilized Zirconia: I. Using Crystal Chemistry to Search for the Ground State on a Glassy Energy Landscape

Yttria-stabilized zirconia (YSZ), a ZrO2-Y2O3 solid solution that contains a large population of oxygen vacancies, is widely used in energy and industrial applications. Past computational studies correctly predicted the anion diffusivity but not the cation diffusivity, which is important for material processing and stability. One of the challenges lies in identifying a plausible configuration akin to the ground state in a glassy landscape. This is unlikely to come from random sampling of even a very large sample space, but the odds are much improved by incorporating packing preferences revealed by a modest sized configurational library established from empirical potential calculations. Ab initio calculations corroborated these preferences, which prove remarkably robust extending to the fifth cation-oxygen shell about 8 Å away. Yet because of frustration there are still rampant violations of packing preferences and charge neutrality in the ground state, and the approach toward it bears a close analogy to glass relaxations. Fast relaxations proceed by fast oxygen movement around cations, while slow relaxations require slow cation diffusion. The latter is necessarily cooperative because of strong coupling imposed by the long-range packing preferences.

cond-mat.mtrl-sci↗
arXiv2017

A Computational Study of Yttria-Stabilized Zirconia: II. Cation Diffusion

Cubic yttria-stabilized zirconia is widely used in industrial electrochemical devices. While its fast oxygen ion diffusion is well understood, why cation diffusion is much slower-its activation energy (~5 eV) is 10 times that of anion diffusion-remains a mystery. Indeed, all previous computational studies predicted more than 5 eV is needed for forming a cation defect, and another 5 eV for moving one. In contrast, our ab initio calculations have correctly predicted the experimentally observed cation diffusivity. We found Schottky pairs are the dominant defects that provide cation vacancies, and their local environments and migrating path are dictated by packing preferences. As a cation exchanges position with a neighboring vacancy, it passes by an empty interstitial site and severely displaces two oxygen neighbors with shortened Zr-O distances. This causes a short-range repulsion against the migrating cation and a long-range disturbance of the surrounding, which explains why cation diffusion is relatively difficult. In comparison, cubic zirconia's migrating oxygen only minimally disturbs neighboring Zr, which explains why it is a fast oxygen conductor.

cond-mat.mtrl-sci↗
arXiv2017

Scalability of Voltage-Controlled Filamentary and Nanometallic Resistance Memories

Much effort has been devoted to device and materials engineering to realize nanoscale resistance random access memory (RRAM) for practical applications, but there still lacks a rational physical basis to be relied on to design scalable devices spanning many length scales. In particular, the critical switching criterion is not clear for RRAM devices in which resistance changes are limited to localized nanoscale filaments that experience concentrated heat, electric current and field. Here, we demonstrate voltage-controlled resistance switching for macro and nano devices in both filamentary RRAM and nanometallic RRAM, the latter switches uniformly and does not require forming. As a result, using a constant current density as the compliance, we have achieved area-scalability for the low resistance state of the filamentary RRAM, and for both the low and high resistance states of the nanometallic RRAM. This finding will help design area-scalable RRAM at the nanoscale.

cond-mat.mes-hall↗