使用集群图注意力网络加速全球大规模银团搜索
Li Fu1, Qiuying Du2, Linwei Sai3
1Key Laboratory of Materials Modification by Laser, Ion and Electron Beams (Dalian University of Technology), Ministry of Education, Dalian 116024, China.
The journal of physical chemistry letters
|August 30, 2024
概括
研究人员利用深度学习和遗传算法发现银的稳定结构 (Agn,30-60个原子). 这种方法解释了Ag48星团的异常稳定性,为原子结构预测提供了更快,更准确的方法.
科学领域:
- 计算化学和材料科学.
- 纳米技术和集群物理.
背景情况:
- 了解银星团稳定性至关重要,传统模型解释了"魔法数字",但并非所有观察到的现象.
- 实验数据表明,银团中具有48个价值电子的稳定性得到了增强,这是现有理论无法完全解释的属性.
研究的目的:
- 使用先进的计算方法确定银团 (Agn,n=30-60) 的全球最小结构.
- 阐明控制这些银团稳定的结构和电子特性.
- 为了解释实验观察到的Ag48集群的增强稳定性.
主要方法:
- 采用深度学习模型,集群图注意力网络 (CGANet),与全面的遗传算法 (CGA) 集成.
- 利用图形处理单元 (GPU) 加速来实现高效的全球结构搜索.
- 计算与密度函数理论 (DFT) 进行了验证,以确保准确性.
主要成果:
- 确定了Agn集群的全球最小结构和代表性异构体 (n=30-60).
- 揭示了竞争的结构图案,包括截断的八面体和二面体,以及更大的群体的二面体基础增长模式.
- 证明了结构性和电子性质的大小依赖的演变解释了Ag48的增强稳定性.
结论:
- 开发的CGANet与CGA相结合,是探索原子潜在能量表面的高效和准确工具.
- 该研究提供了对中型银中结构演变和稳定趋势的全面了解.
- 这些发现成功地合理化了Ag48集群中观察到的令人费解的增强稳定性.
相关概念视频
01:45Vesicular Tubular Clusters
After budding out from the ER membrane, some COPII vesicles lose their coat and fuse with one another to form larger vesicles and interconnected tubules called vesicular tubular clusters or VTCs. These clusters constitute a compartment at the ER-Golgi interface known as ERGIC (Endoplasmic Reticulum Golgi Intermediate Compartment). The ERGIC is a mobile membrane-bound cargo transport system that sorts proteins secreted from ER and delivers them to the Golgi.
With the help of motor proteins such...
With the help of motor proteins such...
01:20Cluster Sampling Method
Appropriate sampling methods ensure that samples are drawn without bias and accurately represent the population. Because measuring the entire population in a study is not practical, researchers use samples to represent the population of interest.
To choose a cluster sample, divide the population into clusters (groups) and then randomly select some of the clusters. All the members from these clusters are in the cluster sample. For example, if you randomly sample four departments from your...
To choose a cluster sample, divide the population into clusters (groups) and then randomly select some of the clusters. All the members from these clusters are in the cluster sample. For example, if you randomly sample four departments from your...