Workflow
密度泛函理论(DFT)
icon
Search documents
DRAM,突破10nm
半导体芯闻· 2025-12-17 10:31
Core Viewpoint - Samsung and its Advanced Technology Research Institute announced a new DRAM technology for manufacturing below 10nm, utilizing a Cell-on-Peri (CoP) architecture that stacks memory cells on top of peripheral circuits, enhancing storage density and efficiency while addressing high-temperature challenges [1][2]. Group 1: Technology Overview - The new CoP architecture differs from traditional methods by placing the surrounding transistors above the memory cells, which mitigates damage during high-temperature stacking processes [1]. - Samsung's technology is named "High-Temperature Resistant Amorphous Oxide Semiconductor Transistor for CoP Vertical Channel DRAM below 10nm" [1]. - The vertical channel transistors, with a channel length of 100nm, can withstand temperatures up to 550 degrees Celsius, maintaining performance stability [2]. Group 2: Performance and Stability - In high-temperature and high-pressure tests, the threshold voltage drift was only -8mV, indicating a stable operational lifespan exceeding 10 years [2]. - The high thermal stability of the transistors is attributed to their ability to suppress ion migration at the channel-electrode interface [2]. - During aging tests, the transistors exhibited minimal degradation in drain current, demonstrating robust performance [2]. Group 3: Development Stage and Future Applications - The technology is currently in the research and development phase and is expected to be applied in future DRAM processes below 10nm (0a and 0b nodes) [2].
David Baker最新Nature论文:AI从头设计金属水解酶,无需实验优化,催化效率提升千倍
生物世界· 2025-12-04 08:30
撰文丨王聪 编辑丨王多鱼 酶的从头设计,旨在构建含有理想活性位点的蛋白质,这些位点周围的催化氨基酸残基能够稳定目标化学 反应的过渡态。此前已有研究利用蛋白质从头设计来生成新的 金属水解酶 ( Metallohydrolase ) , 但这 些酶的活性和效率相对较低, 需要经过大量的定向进化才能达到天然酶的活性和效率水平。 排版丨水成文 David Baker 教授 David Baker 团队之前开发的用于蛋白质从头设计的生成式 AI 工具—— RFdiffusion,可以解决上述难 题,但其需要为每个催化氨基酸残基指定序列位置和主链坐标,这限制了 设计空间范围。 2025 年 12 月 3 日,诺奖得主、蛋白质设计先驱 David Baker 教授团队在国际顶尖学术期刊 Nature 上发 表了题为 : Computational design of metallohydrolases 的研究论文。 该研究利用新一代 AI 蛋白质设计工具—— RFdiffusion2 , 成功设计了活性极高的锌金属水解酶,其催化 效率比之前设计的金属水解酶高出上千倍 。更令人惊叹的是,这些高性能酶完全"从头开始"设计,且无 ...
新宙邦CBS添加剂登上《Small》:高压电解液难题的新解法
高工锂电· 2025-08-26 11:01
Core Viewpoint - The battery industry is undergoing a new round of material upgrades driven by the demands of electric vehicles for lightweight and ultra-fast charging, as well as consumer electronics for thinness and long battery life. High-performance materials such as high-nickel cathodes and silicon-carbon anodes are accelerating their application [2]. Group 1: High-Performance Materials - High-voltage cathodes (4.5 V–4.7 V) are becoming a key focus in the industry due to their ability to enhance energy density and fast charging capabilities. However, issues such as electrolyte oxidation and transition metal ion dissolution pose challenges to battery lifespan and safety [2][4]. - The research team has developed a hybrid molecule CBS (Carbonate Bis(Sulfate)) that combines the structural advantages of carbonates and sulfates, addressing the critical need for a stable electrode-electrolyte interface [4][5]. Group 2: Performance Improvements - The CBS additive significantly enhances battery lifespan, with a capacity retention rate of 94% after 600 cycles at 25°C, and 90% after 1000 cycles at 45°C, showing a much lower impedance growth compared to the baseline [6]. - At 60°C, batteries with CBS maintained a capacity retention rate of 90% after 30 days, with only 3% volume expansion, while the baseline electrolyte showed only 13% capacity retention and 34% volume expansion [7]. - Safety is greatly improved, with the decomposition temperature of the SEI formed by CBS rising from 124.8°C to 140°C, and the CEI decomposition temperature increasing from 93.1°C to 121.4°C, effectively delaying thermal runaway risks [8]. Group 3: Versatility Across Battery Systems - CBS shows advantages beyond the NCM system, performing well in other mainstream battery chemistries. For instance, in the LiCoO₂ system, the capacity retention rate exceeds 80% after 800 cycles at 25°C, compared to only 37% for the baseline [11]. - In the LiMn₀.₆Fe₀.₄PO₄ system, CBS effectively inhibits Mn²+ dissolution at high temperatures, demonstrating superior cycling stability compared to baseline and PS electrolytes [12]. Group 4: Research and Collaboration - The research on CBS was led by a team from Shenzhen New Zobang Technology Co., Ltd. and Southern University of Science and Technology, highlighting a strong collaborative effort in advancing battery chemistry and materials [16]. - Since the release of the CBS technology, multiple leading industry players have shown interest, and the technology is being widely applied in the ICT industry. The company is also focusing on intellectual property protection and value conversion [17].
史上引用量第一的论文,是哪篇?
Hu Xiu· 2025-07-20 11:55
Core Insights - The most cited research paper of all time is a 1951 paper published in the Journal of Biological Chemistry, detailing a method for determining protein content in solutions, with over 350,000 citations as of the latest statistics [1][4][10] - The owner of the Web of Science (WoS) database, Clarivate, reported that a significant portion of highly cited papers are related to biological experimental techniques, with the threshold for entering the top 100 citations rising from approximately 12,000 in 2014 to over 30,000 now [2][7] - An analysis of other databases, Dimensions and OpenAlex, shows slight variations in rankings, but the 1951 paper remains at the top across all platforms [3][4] Citation Trends - There are only three papers with over 200,000 citations, all related to biological experimental techniques, and the top three papers in the new ranking remain unchanged from 2014 [7] - Sixteen papers published in the 21st century have entered the historical top 50, indicating a significant rise in citations for software descriptions and computer-aided research results [8] - The increase in annual publication rates and the rise of online research and social media exposure are contributing factors to the rapid citation growth of newer papers [9] Future Projections - If the current citation growth rates continue, certain recent papers, such as those on AI and density functional theory (DFT), may surpass the 1951 paper by 2030 [10]