Dense Stacks Research Articles

Thermal boundary conductance enhancement of the Si/diamond interface via atomic transition strategy

Developing multi-chip systems introduces significant thermal management challenges, due to dense vertical stacking and interconnections. Specifically, the increased number of interfaces within the heat conduction path of the chips significantly impedes the effective heat transfer. This study utilizes a series of molecular dynamics simulations to explore how the structure and atomic composition at interfaces affect their thermal conduction abilities. This study initially reveals substantial differences in thermal conduction capabilities between Si/Diamond, SiC/Diamond, and Diamond/Diamond interfaces. Further investigations focus on interfaces between different structures of SiC and diamond, clearly identifying the atomic composition and structure at the interface as key factors influencing thermal boundary conductance (TBC). Based on these findings, the study proposes a strategy for atomic transition that involves inserting SiC into the Si/Diamond interface. Under optimal thickness conditions for the SiC transition layer, a significant increase in theoretical TBC is achieved, from 477.1 MW/m²K to 701.1 MW/m²K, which is twice the value predicted by the existing diffusive mismatch model (DMM) for Si/Diamond interfaces. Lastly, through the developed 3D IC model, the study examines the impact of TBC variations on the peak temperature of the entire device, thereby further emphasizing the importance of enhancing interface thermal conduction capabilities. This series provides strategic guidance for thermal management in 3D ICs and offers a theoretical basis for chip design and material selection.

A Versatile Method for Preparation of BrCOFs Aerogels and Efficient Functionalization via Suzuki-Miyaura Reaction.

Covalent organic frameworks (COFs) aerogels solve the restrictions on processability and application caused by the insolubility and non-fusibility of powders while avoiding the inaccessibility of pore structures by dense stacking. At the current start-up stage where COFs aerogels are scarce and difficult to synthesize, design of generalized synthetic methods play an indispensable role in guiding and developing COFs aerogels. Moreover, evolving the functionality of COF aerogels is equal vital, which achieves higher performance and broader practical applications. In this work, for the first time, processable BrCOFs aerogels have been synthesized without vacuum by seven kind polar solvents, which realizes general preparation of BrCOFs aerogels. It is extremely friendly to the inapplicability for some scenarios. Furthermore, by Suzuki-Miyaura cross-coupling reaction, BrCOFs aerogels are endows with cyano groups (-CN), trifluoromethyl (-CF3) and methyl sulfonyl (-SO2-CH3) efficiently. As a proof-of-concept, BrCOFs-SO2-CH3 aerogels served as a quasi-solid electrolyte for lithium-metal batteries (LMBs), which effectively enhance the performance of batteries.

Biomineralization Inspired the Construction of Dense Spherical Stacks for Dendrite-Free Zinc Anodes.

Aqueous zinc-ion batteries (AZIBs) are considered to be one of the most promising energy storage systems due to their high degree of safety and low cost. However, the deposition of the uncontrolled zinc dendritic morphology on the surface of the Zn anode seriously reduces the cycle life of AZIBs. Herein, inspired by natural biomineralization, a uniform spherical zinc deposition is achieved via the addition of a biological macromolecule to the electrolyte. The proposed biological macromolecule makes Zn2+ undergo dense spherical deposition by regulating the relative growth rate of high-index planes of metal zinc, effectively avoiding the destruction from irregular zinc dendrites. The symmetric cell with the addition of such a biological macromolecule can be stably cycled for >3200 h at 1 mA cm-2 for 1 mAh cm-2. This work sheds light on expanding morphological regulation of metal deposition to spherical shapes for stable metal anodes in addition to a single-crystal plane control strategy.

Superior Charge Density of Triboelectric Nanogenerator via Trap Engineering

AbstractTriboelectric nanogenerator (TENG) offers a novel approach for converting high‐entropy mechanical energy into electrical energy, yet achieving high charge density remains critical. Optimizations using dielectrics with high specific capacitance have mitigated air breakdown, but charge loss within dielectrics persists as a limiting factor. Here, based on poly(vinylidene fluoride–trifluoroethylene–chlorofluoroethylene) (P(VDF‐TrFE‐CFE)) with high specific capacitance, (P(VDF‐TrFE‐CFE)) composites’ trap density and energy are engineered using high‐polarity interfaces from barium titanate (BTO) nanoparticles and dense chain segment stacking induced by electrostatic interaction with polyetherimide (PEI) to enhance charge retention capability. With modified high interfacial traps, an ultrahigh charge density of 9.23 mC m−2 is achieved in external charge excitation (ECE) TENG using 0.2 vol% PEI/P(VDF‐TrFE‐CFE) film, marking the highest charge density reported for single‐unit TENGs. This work provides novel material strategies for high‐performance TENGs, paving the way for their large‐scale practical applications.

Stress-driven triple junction reconstruction facilitates cooperative grain boundary deformation

Triple junctions (TJs), essential components linking neighboring grain boundaries (GBs), are of great significance for the deformation of entire GB networks in polycrystalline materials. However, kinetic behaviors of TJs and their coupling with GB plasticity remain largely unexplored, especially at atomic scale. Using atomistic in situ nanomechanical testing, we reveal a regime of dynamic TJ reconstruction for accommodating the coordinated deformation of GB network in gold and platinum polycrystals, proceeding through different modes of structural transformations, including disordered atomic arrangement, subgrain, dense stacking faults, and nanotwins. Such TJ reconstruction preferentially nucleates at TJs predicted with strong dragging effect, which serves as an effective route to facilitate the cooperative motion of neighboring GBs, in contrast to the widely-believed TJ deformation in steady state. This reconstruction-coordinated TJ kinetics provides novel insights into complicated GB network evolution and calls for a revisit of TJ roles in polycrystalline materials.

Single‐Luminophore Molecular Engineering for Rapidly Phototunable Solid‐State Luminescence

AbstractSmart materials enabling emission intensity or wavelength tuning by light stimulus have attracted attention in cutting‐edge fields. However, due to the general limitation of the dense molecular stacking (in solid states, especially in crystals) on photoresponsivity, constructing rapidly phototunable solid‐state luminescent systems remains challenging. Herein, we present a new luminophore that serves as both a photoresponsive and a luminous group with enhanced conformational freedom to attain this goal, namely, relying on photoexcitation‐induced molecular conformational change of an ionized persulfurated arene based on weak intermolecular aliphatic C−H⋅⋅⋅π interaction. Together with the phosphorescence characteristic of the molecule itself, rapidly enhanced phosphorescence upon irradiation can be observed in a series of phase states, like solution state, crystal, and amorphous state, especially with a high photoresponsive rate of 0.033 s−1 in crystal state that is superior to the relevant reported cases. Moreover, a rapidly phototunable afterglow effect is achieved by extending this molecule into some polymer‐based doping systems, endowing the system with unique dynamic imaging and fast photopatterning capabilities. This single‐luminophore molecular engineering and underlying mechanism have implications for building other condensed functional materials, principally for those with stimuli responses in solid states.

Boosting Organic Solar Cells to Over 18 % Efficiency through Dipole-Dipole Interactions in Fluorinated Nonfused Ring Electron Acceptors.

This study successfully designed and synthesized two nonfused ring electron acceptors, 412-6F and 412-6Cl, modified with fluorine and chlorine substituents, respectively. Single-crystal analysis revealed that 412-6F possesses a planar molecular backbone and exhibits pronounced dipole-dipole interactions between the fluorine atoms on the lateral phenyl groups and the carbonyl oxygen atoms on the end groups. This specific interaction promotes dense end-group stacking, leading to a reduced interlayer spacing. Improved crystallinity and coherence length are observed in the D18 : 412-6F blend film. Conversely, 412-6Cl adopts a more distorted configuration and lacks these interactions. As a result, the organic solar cell (OSC) based on D18 : 412-6F achieved a remarkable power conversion efficiency of 18.03 %, surpassing the performance of the D18 : 412-6Cl OSC. This underscores the importance of designing novel acceptors with beneficial intermolecular interactions to enhance OSC efficiency, thus providing a new direction for organic photovoltaic advancement.

Boosting Zinc‐Ion Storage Capability in Longitudinally Aligned MXene Arrays with Microchannel Architecture

AbstractFlexible Zn2+ ion hybrid capacitors (ZHCs) will play a crucial role in developing next‐generation wearable products, which demand portability, durability, and environmental adaptability. To further meet these requirements, Ti3C2Tx MXene with exceptional conductivity and robust mechanical properties can be utilized as cathodes, except for challenges such as the dense stacking of Ti3C2Tx nanosheets. In this study, a novel MXene cathode architecture has been developed with the facilitation of an ice template, which creates longitudinally aligned Ti3C2Tx arrays with microchannels. The introduction of hydrochloric acid to the Ti3C2Tx slurry induces a crumpled morphology, increasing active sites and enhancing ion transport with expanded interlayer spacing. Interestingly, experimental results and COMSOL simulations verify that the cathode structure also has effective impacts on the Zn anode with a weakened ion concentration gradient and suppressed dendrite formation. Consequently, the ZHCs exhibit enhanced electrochemical performance with excellent rate performance and long‐term cycling stability (enduring over 50 000 cycles at 5 A g−1) and further deliver a reliable low‐temperature operation by applying an anti‐freezing electrolyte. Moreover, flexible ZHCs assembled with gel electrolytes demonstrate excellent flexibility, improved rate performance, and mechanical stability, making them well‐suited for flexible electronics that power flexible LED panels.

A polyurea capturing both high strength and rapid self-healing tailored for intelligent barrier anti-corrosion system

The long-term repair of organic coatings in underwater environments is hindered by microcracks and their poor healing performance. In this study, a high-strength and rapidly self-healing polyurea (PU3) coating that combines asymmetric dense stacking of supramolecular units with mismatched rigid and flexible polymer chains was developed. By exploiting the interaction between a 2D sheet-like Ti3C2TX MXene and finely tailored hydrogen-bonded structures of PU3, a smart biomimetic anticorrosion composite coating (PU3-1.0 wt% MXene) was produced. This coating achieved rapid (15 s), specific, and repeatable in situ repair underwater when exposed to near infrared laser light. Electrochemical tests confirmed its excellent anticorrosion capability.

Polyphenol assisted assembly of prussian blue analogs nanocrystals decorated 3D GO functional microspheres for polymer fiber bifunctional membrane with high oil-water emulsion permeability and powerful photocatalytic self-cleaning ability

Due to the intricate components of wastewater and physical stacking of graphene oxide (GO), the current GO membranes exhibit the fatal defects of low porosity and poor anti-fouling ability. Herein, a three dimensional (3D) FeCo-PBA@GO functional microsphere was constructed via anchoring GO nanosheets onto hard template of 3D polymethyl methacrylate (PMMA) microsphere driven by hydrogen bond and van der Waals force. On this basis, the Prussian blue analogs nanocrystals (FeCo-PBA) grew in situ on such functional microsphere and further assembled onto porous polyacrylonitrile (PAN) fiber mat through the tannic acid assisted spraying technique, resulting in highly permeable and photocatalytic self-cleaning FeCo-PBA@GO/PAN fibrous composite membrane. Interestingly, the unique 3D nanostructure of FeCo-PBA@GO microsphere arranging on membrane surface effectively prevented the dense physical stacking of GO nanosheets and increased the oil repellence property, thereby regulating the water channel and breaking through permeability limitation of membrane. As a result, the composite membrane exhibited superior separation fluxes (˃3029 L m−2 h−1) for different oil/water (O/W) emulsions, with rejection rate above 99.0 %. Owing to rich active sites and enhanced electron/mass transport property of FeCo-PBA functional layer, the composite membrane could effectively degrade various dyes by peroxymonosulfate (PMS) activation, with the degradation efficiency of 93.1–96.9 % within 40 min and demonstrated favorable antifouling and photocatalytic self-cleaning ability after cycling separation of O/W emulsion. In addition, the generated multiple reactive oxygen species and photocatalytic mechanism were proposed. This work provides a versatile pathway to prepare multifunctional GO advanced membranes for large-scale treatment of complex emulsified oily wastewater.

Gradient Hierarchically Porous Ionic-Junction Fibers of Wet-Spun Carboxymethyl Cellulose Coagulated with Copper Sulfate.

Polyelectrolyte-based ionic-junction fibers newly serve as signal transmission and translation media between electronic devices and biological systems, facilitating ion transport within organic matrices. In this work, we fabricated gel filaments of carboxymethyl cellulose (CMC) chelated with Cu(II) ions through wet-spinning, using a saturated coagulant of CuSO4. Interestingly, the as-spun fibers exhibited dramatic 3D porous frameworks that varied with the temperature and precursor concentration. At 20 °C, the Cu(II) chelation networks favored the formation of well-organized cellular chambers or corrugated channels, displaying dense stacking patterns. However, critical transitions from cellular chambers to corrugated channels occurred at precursor dope concentrations of approximately 2 and 7 wt %, with the porous structure diminishing beyond 8 wt %. We have proposed schematic diagrams to mimic the 3D pore structure, dense porous stacking, and formation mechanism, according to electronic micrographs. Our investigations revealed that the distinct ion-junction channels or chambers are under the control of axial drawing extension as well as the outside-inside penetration of Cu(II) ions into the dope and inside-outside diffusion of water into coagulants. Therefore, controlling the metal chelation-water diffusion process at specific temperatures and concentrations will offer valuable insights for tailoring ionic-junction soft filaments with gradient hierarchically porous structures and shape memory properties.

Enhanced interlayer interaction in sulfonated CONs membrane by amino-rich CONs enabling ultrafast proton transport

Covalent organic framework nanosheets (CONs) with porous crystalline features and ultrathin thickness are ideal candidates as membrane building blocks to form well-defined transfer nanochannels. The formidable challenge behind self-supporting CONs membrane lies in weak non-covalent interlayer interactions and thus loose stacking, insufficient strength and structure stabilities. Herein, we propose the fabrication of interlayer force-strengthened freestanding CONs membrane through the electrostatic attraction bridge effect of positively-charged amino-rich CONs (CON–NH2) to negatively-charged sulfonated CONs (CON–SO3H). Ultrathin and large lateral sized CON–SO3H and CON–NH2 are synthesized, followed by restacking to prepare freestanding CONs membrane with CON–SO3H as the membrane bulk. Benefiting from effective interlayer interconnection due to strong electrostatic interaction, the obtained CON–SO3H/CON–NH2 membrane displays features of ultrahigh integrity, dense stacking, eminent water/acid/base/organic solvents stabilities and mechanical strength (109 MPa). The shortened –SO3H distance contributes to construct site-continuous transfer pathways, and the deprotonated –SO3H and protonated –NH2 form acid–base pairs to decrease interfacial resistance, which impart membrane superior proton conductivity of 486 mS cm−1 (80 °C, 100% RH). This interlayer force enhancement strategy offers a promising perspective on achieving densely-stacked CONs membrane with ultrahigh mechanical property and conduction performance for fuel cell application.

Sub-nano porous graphene-based membranes enabled by in-situ-grown ZIF-8 for enhanced CO2 capture

Graphene oxide (GO) membranes hold promise for selective CO2 adsorption over N2 due to their carboxyl groups, making them attractive for flue gas CO2 capture. However, challenges such as limited carboxyl content, flexible nanosheets, and dense layer stacking hinder their N2/CO2 gas selectivity, pressure resistance, and permeability. In this study, we introduce a specific sub-nanometer framework structure within GO membranes by in-situ growing ZIF-8 nanocrystals. This novel approach, achieved through simultaneous infiltration of zinc nitrate and 2-methylimidazole precursors on both sides of the membrane, enhances CO2 capture and pressure resistance capabilities of resulting ZIF-8@GO composite membranes. Additionally, the incorporation of carboxylated wrinkled graphene (WG) and ethylenediamine (EDA) cross-linking agent molecules improves CO2 capture capability and introduces reinforced porous structures to enhance permeability. Experimental results demonstrate remarkable permeability, selectivity, and pressure resistance of the prepared ZIF-8@EDA-GO/WG composite membranes. Under a gas pressure of 0.2 MPa, permeability reaches 1850 GPU, with theoretical selectivity for N2/CO2 single gas and separation factors for mixed gases of 18.3 and 32.3, respectively, surpassing conventional GO membranes (3 GPU, 1.1, and 1.2). Even under elevated air pressure conditions (1.2 MPa), the composite membranes maintain a theoretical selectivity of 13.4. Our CO2 capture membrane design strategy not only advances the development of functional membranes but also contributes to mitigating CO2 emissions from flue gas, thereby combating global warming.

Supramolecular assemblies derived from methyl-substituted cucurbit[5]uril and lanthanide nitrates

Interaction of the lanthanide nitrates M(NO3)3 (M = Gd, Eu) with methylcucurbit[5]uril (Me10Q[5]) in the presence of transition metal chlorides (ZnCl2 and FeCl3) in acidic media resulted in the isolation of the complexes [Me10Q[5]Gd(H2O)2Cl Gd(H2O)6](ZnCl4)2∙Cl∙8.9H2O (1) and [Me10Q[5]Eu(H2O)3Cl(H3O)](FeCl4)3 (2). The molecular structures of 1 and 2 have been determined by single crystal X-ray crystallography, and reveal discrete complexes which are involved in dense stacking with adjacent Me10Q[5]s linked via H-bonding and/or metal anions resulting in a supramolecular assembly.

Valorization of magnesium slag and CO2 towards a low carbon fiber reinforced cement board

Pyrometallurgical extraction of magnesium from dolomite ores results in a large amount of magnesium slag discharge and CO2 emission, posing severe environmental hazards. This study introduces a new approach for the simultaneous upcycling of magnesium slag and CO2 by taking advantage of the high carbonation reactivity of magnesium slag. A CO2-solidified magnesium slag-based fiber cement board with high-strength and high-toughness (CHFCB) is developed, using magnesium slag as the binder, CO2 as the activator, and a mix of pulp and PAN fibers as the reinforcement. The optimal CHFCB preparation process was explored by adjusting the fiber ratio, molding pressure, and CO2 curing system, and the mechanisms underlying its strength and toughness were analyzed. The results indicated that with the optimal preparation, CHFCB exhibited a water-saturated flexural strength of nearly 20 MPa and a drying flexural strength of over 25 MPa. The ultimate deflection could be increased by up to 80 % or more during the bending deformation of CHFCB, while the tensile strain could be improved by over 50 % in the tensile fracture process of CHFCB. After 100 freeze-thaw cycles, the water-saturated flexural strength ratio of CHFCB reached 97.2 %, showing excellent durability. The microanalysis revealed that the dense stacking of calcite-type calcium carbonate and the presence of PAN fibers contributed to the high-strength and high-toughness of CHFCB. Additionally, the environmental and economic benefits of CHFCB were assessed, affirming its role as a sustainable building material that promotes energy efficiency and a green economy.

Predictive modelling and identification of key risk factors for stroke using machine learning

Strokes are a leading global cause of mortality, underscoring the need for early detection and prevention strategies. However, addressing hidden risk factors and achieving accurate prediction become particularly challenging in the presence of imbalanced and missing data. This study encompasses three imputation techniques to deal with missing data. To tackle data imbalance, it employs the synthetic minority oversampling technique (SMOTE). The study initiates with a baseline model and subsequently employs an extensive range of advanced models. This study thoroughly evaluates the performance of these models by employing k-fold cross-validation on various imbalanced and balanced datasets. The findings reveal that age, body mass index (BMI), average glucose level, heart disease, hypertension, and marital status are the most influential features in predicting strokes. Furthermore, a Dense Stacking Ensemble (DSE) model is built upon previous advanced models after fine-tuning, with the best-performing model as a meta-classifier. The DSE model demonstrated over 96% accuracy across diverse datasets, with an AUC score of 83.94% on imbalanced imputed dataset and 98.92% on balanced one. This research underscores the remarkable performance of the DSE model, compared to the previous research on the same dataset. It highlights the model's potential for early stroke detection to improve patient outcomes.

Effect of SiC particle content on the mechanical behavior and deformation mechanism of SiCp/Al composite under high-frequency dynamic loading

SiCp/Al composites exhibit excellent structural properties but suffer from inherent brittleness, which has brought rigorous challenges on the materials processing. Ultrasonic-vibration-assisted approach offers a promising solution to improve the processability. However, the underlying mechanisms of mechanical response and defect suppression under ultrasonic action have been rarely studied for SiCp/Al composites, hindering the development of high-efficiency processing. This work aims to investigate the mechanical behavior and microstructural evolution in this class of composites with varying SiC volume fractions up to 45% under high-frequency dynamic load induced by ultrasonic vibration. The results showed that the deformation stress of SiCp/Al composites under ultrasonic compression was significantly reduced to only about 1/10 of that under conventional quasi-static loading. Moreover, as the SiC fractions increasing, this ultrasonic softening effect of SiCp/Al became more pronounced. The plastic limit of SiCp/Al with 45% SiC fraction nearly doubled under ultrasonic compression without any macroscopic cracks. The subsequent nanoindentation tests revealed a transition in work-hardening behavior of SiCp/Al with different SiC fractions under ultrasonic loading. Microstructure characterization was conducted to analyze the grain refinement, recrystallization, and deformation textures within the deformed areas, revealing the deformation mechanism of SiCp/Al under different loading mode. TEM observations confirmed the distribution of dislocations in Al matrix and dense stacking faults in SiC particles after ultrasonic compression, indicating that the plastic behaviors were enhanced in both Al matrix and SiC phase under the high-frequency cyclic stress. This research revealed the distinct deformation mechanisms of SiCp/Al composites under high-frequency dynamic loading, exhibiting extremely low deformation stress accompanied with significantly enhanced plasticity. The results provide theoretical insights for the high-efficiency processing and rapid property regulation of particle-reinforced-composites based on ultrasonic-vibration-assisted technologies.

Synergistic improvement in gas separation performance of MMMs by porogenic action and strong molecular forces of ZIF-93

In the present study, ZIF-93 metal–organic frameworks were prepared. N2 adsorption tests revealed that the particle size, specific surface area, and pore size of ZIF-93 were 30 nm–45 nm, 869.40 m2/g, and 0.36 nm, respectively. Pebax-based mixed matrix membranes (MMMs) were prepared using ZIF-93 as the filler. The pore diameters of ZIF-93 lied between the kinetic diameters of CO2 and N2; thus, its pore structure provided a fast transfer channel for CO2 and increased the permeability of the membrane. Further, X-ray diffraction showed that the membrane became denser upon introducing ZIF-93 because of the formation of a bridging structure between the Pebax molecular chains and ZIF-93. A denser stacking by moving the Pebax molecular chains closer to each other effectively bettered the CO2 selectivity of the mixed matrix membrane. The test results showed that when the ZIF-93 loading was 10 wt%, the CO2 permeability was 84.85 barrer and the CO2/N2 selectivity was 65.76 at 0.4 MPa, which are 51.57 % and 65.50 % higher than those of the pure Pebax membrane, respectively. When the feed pressure was increased to 0.8 MPa, the CO2 permeability was 94.5 barrer and the CO2/N2 selectivity is 87.5, which are 51.57 % and 65.50 % higher compared with those of the pure Pebax membrane, in order. The membrane performance significantly exceeded that of the 2008 Robeson curve.

H3DM: A High-bandwidth High-capacity Hybrid 3D Memory Design for GPUs

Graphics Processing Units (GPUs) are widely used for modern applications with huge data sizes. However, the performance benefit of GPUs is limited by their memory capacity and bandwidth. Although GPU vendors improve memory capacity and bandwidth using 3D memory technology (HBM), many important workloads with terabytes of data still cannot fit in the provided capacity and are bound by the provided bandwidth. With a limited GPU memory capacity, programmers should handle the data movement between GPU and host memories by themselves, causing a significant programming burden. To improve programming ease, GPUs use a unified address space with the host that allows over-subscribing GPU memory, but this approach is not effective in terms of performance once GPUs encounter memory page faults. Many recent works have tried to remedy capacity and bandwidth bottlenecks using dense non-volatile memories (NVMs) and true-3D stacking. However, these works mainly focus on one bottleneck or do not provide a scalable solution that fits future requirements. In this paper, we investigate true-3D stacking of dense, low-power, and refresh-free non-volatile phase change memory (PCM) on top of state-of-the-art GPU configurations to provide higher capacity and bandwidth within the available area and power budget. The higher density and lower power consumption of PCM provide higher capacity through integrating more cells in each 3D layer and enabling stacking more layers. However, we observe that stacking more than six layers of pure-PCM memory violates the thermal constraint and severely harms the performance and power efficiency due to its higher write latency and energy. Further, it degrades the lifetime of GPU to less than one year. Utilizing a hybrid architecture that leverages the benefits of both DRAM and PCM memories has been widely studied by prior proposals; however, true-3D integration of such a hybrid memory architecture especially on top of state-of-the-art powerful GPU architecture has not been investigated yet. We experimentally demonstrate that by covering 80% of write requests in DRAM and eliminating refresh overhead, true-3D stacking of eight 32GB layers of PCM along with two 8GB layers of DRAM is possible resulting in a total of 272GB memory capacity. Based on the explored design requirements, We propose a 3D high-bandwidth high-capacity hybrid memory (H3DM) system utilizing a hybrid-3D (H3D)-aware remapping scheme to reduce expensive PCM writes to under 20% while avoiding DRAM refresh overhead. H3DM improves the performance up to 291% compared to the baseline GPU architecture while remaining within only 3% of an ideal case with DRAM-like access latency, on average. Moreover, by increasing the dataset size above the baseline GPU memory space, H3DM improves performance and power up to 648% and 87% compared to the baseline GPU architecture since it avoids expensive data transfers through off-chip communication links.

Dense Stacking of Millimeter‐Sized Perovskite Single Crystals for Sensitive and Low‐Bias Hard X‐Ray Detection

AbstractMetal halide perovskite single crystals are promising for hard X‐ray detection, but growth of large‐sized thin single crystals and inhibition of halide ion migration under high bias are challenging. Herein, the preparation of dense stacking of oriented millimeter‐sized perovskite single crystals with capability of large‐area fabrication, controlled thickness, and high ion migration activation energy (Ea)is reported. The oriented growth results in absence of grain boundaries parallel with the substrates, leading to large carrier mobility‐lifetime (µτ) product of 2.7 × 10−3 cm2 V−1, which is comparable to values of many perovskite bulk single crystals. Under low bias, hard X‐ray detectors exhibit high sensitivity of 1.2 × 104 µC Gy−1 cm−2 and low detection limit of 87.5 nGy s−1, which are comparable to those of many single‐crystal hard X‐ray detectors under high bias. Moreover, combination of low bias and high Ea results in weak halide ion migration and small dark current drift of 3.2 × 10−4 nA cm−1 s−1 v−1. Benefited from the superior detector performance, high‐contrast hard X‐ray imaging can be obtained at a low dose rate of 1.26 µGy s−1. The work may promote the application of perovskite hard X‐ray detectors in practical imaging area.