Tag
MIT alumni profile of Phoenix Tailings, cofounded by Tomás Villalón Jr., which uses a proprietary electrochemical process to extract rare earth metals from mining waste without generating toxins or carbon emissions.
MIT engineers developed a recyclable thermoplastic yarn made from two polyethylene variants that matches the strength and stretch of spandex-based yarns and can be melted and respun for up to 10 cycles without losing performance, potentially enabling a circular textile economy.
MIT researchers published a Nature Communications paper on a multimodal AI model that combines 3D atomic structures with natural language to better predict and design crystalline materials, outperforming text-only models in identifying promising candidates like solid-state lithium-ion battery electrolytes.
O-kRystal, a periodic E(3)-equivariant graph neural network, learns corrections to electronic states beyond DFT, achieving HSE06 band predictions across nearly 6,000 materials, G0W0 quasiparticle energies in 2D systems, and a 68.7% reduction in band energy MAE on strongly correlated materials.
Charles C. Mann explores the high-tech revival of porcelain, from its ancient Chinese origins in kaolin clay to modern ultra-compact surfaces like Dekton, arguing that materials science is transforming the built environment as profoundly as the digital revolution.
Scientific analysis and a two-year, long-term SEM experiment on whether liquid metal thermal interface materials “dry out” on bare copper surfaces: Gallium reacts with copper to form the intermetallic compound CuGa₂ (with a thermal conductivity of approximately 100 W/(m·K) and a growth rate of about 1 micrometer per day). Although nickel plating forms NiGa₅, the layer is approximately one-tenth as thick, so nickel plating remains the better choice; meanwhile, a gallium supply crisis has driven prices up nearly tenfold in five years.
希伯来大学研究人员受二战Pykrete启发，通过在冰中添加植物纤维素纳米晶体和定制抗冻蛋白，创造出强度是普通冰10倍、韧性提升70倍的"BioPykrete"生物冰材料，有望用于极地等偏远地区的低碳可持续建筑。
Researchers developed a chemistry-based approach that allows new tattoos to be turned on and off, effectively creating switchable, reconfigurable body art. The work highlights advances in responsive materials for wearable applications.
ASPIRE 是一个通过等变集合预测器 Ev-Quiformer 预测原子环境的多个鞍点候选、再用 Dimer 搜索进行物理精修的框架，用于加速辐照材料中的扩散与缺陷演化事件发现，在 BccFe 四空位基准上以约一半的力评估开销达到 77.20% 的参考事件覆盖率，并发布了两个新数据集。
PXtal 提出一种在模态间信息不对称条件下对齐粉末X射线衍射（PXRD）与晶体结构表示的框架，利用非平衡最优传输（UOT）和广义Kullback-Leibler散度监督传输计划，在PXRD到晶体候选检索及下游材料任务上优于基线模型。
Scientists studying Chernobyl 'lava' particles have found the nuclear fuel has remained unexpectedly stable over 40 years, with implications for long-term radioactive contamination management.
Researchers developed a light-activated hydrogel that can evolve alongside living tissue, potentially enabling more flexible brain-computer interfaces and soft robots that adapt to biological environments.
Vals AI 研究称由 Claude Opus 5.5 主导的 AI 团队从一篇1999年被遗忘的论文中重新挖掘出一种候选材料，可能将计算机内存速度提升1000倍，展示了 AI 在科研发现上的新潜力。
论文提出GanJiang——一个用于粉末X射线衍射分析的自学习智能体，基于XMatcher、XQueryer、XDecomposer和WPEM构成的衍射分析生态系统，通过诊断失败并修订可执行技能来积累经验，无需重训语言模型即可在单相和多相识别基准上取得领先性能。
An article discussing how artificial intelligence is poised to transform materials science, accelerating the discovery and design of new materials through ML-driven approaches.
Vals AI reports that a team of AI agents (Claude Opus 5.5) designed one and identified an existing candidate room-temperature antiferromagnetic semiconductor material, predicted by calculations to have the properties needed for next-generation spintronic computer memory. The work maps out the trade-offs between ferromagnetic, antiferromagnetic, and Luttinger-compensated materials for storage devices.
A 2025 review or study examining how plant-based materials can replace petroleum-based products, covering sustainable alternatives in materials and chemical feedstocks.
Rice University researchers introduce CompMat-Bench, a 94-task benchmark derived from published computational materials studies that evaluates AI agents on preparing inputs and analyzing outputs of expensive simulations, avoiding costly reruns during grading. Agents based on three LLMs achieve 66.0–90.4% pass rates with full guidance, but performance drops under longer workflows or reduced methodological hints, with most failures stemming from scientific errors rather than software usage.
A lighthearted or investigative look at impurities in materials or samples, likely relating to semiconductor, quantum, or lab research findings — a quirky take on contamination, detection, or quality-control stories in tech and science.
CARAT introduces a matched-evidence benchmark to test whether materials LLMs truly reason over crystal structures or merely recite answers printed in their inputs, finding large apparent gains that largely measure baseline gaps rather than reasoning ability.