Configurational constraints on glass formation in the liquid calcium aluminate system
Configurational constraints on glass formation in the liquid calcium aluminate system
复制标题
DOI：
10.1088/1742-5468/ab47fc
复制
发表时间：
2019-10-01
影响因子：
2.4
通讯作者：
Hennet, Louis
中科院分区：
文献类型：
--
作者：
Drewitt, James W. E.;Jahn, Sandro;Hennet, Louis
我们报告新的时间分辨同步辐射X射线衍射（SXRD）测量跟踪结构转变铝酸钙（CaO）（x）（Al 2 O3）（1？x）玻璃形成过程中的液体，并回顾了中子衍射与同位素替代（NDIS）实验的最新进展，结合非球面离子模型分子动力学（AIM-MD）模拟，以确定原子尺度的结构约束的玻璃形成能力。时间分辨的测量揭示了在过冷过程中发生的短期和中期订购的实质性变化。在等摩尔组成中，x？？=？？0.5（CA），液体经历了一个显着的结构重组的玻璃化作为过协调的AlO 5多面体和氧三簇分解，形成一个网络的主要角共享AlO 4四面体。这是伴随着形成的支链的边缘和面对共享钙为中心的CaOy？多面体有助于阳离子有序的中间长度尺度。富钙端员的玻璃形成系统x？？=？？0.75（C3 A）主要由AlO 4四面体组成，但存在%未连接的AlO 4单体和Al 2 O 7二聚体，这代表了一个阈值，在该阈值之后，玻璃不再支持形成无限连接的网络。总的来说，AIM-MD模拟与SXRD和NDIS实验非常一致，表明了准确的势模型。然而，模拟的玻璃结构和实验测量值之间的小差异是明显的，这表明在模拟的玻璃轨迹中存在小程度的液体状有序。这可能是由于短模拟时间尺度，这是不代表的粘性动力学过程中涉及过冷和玻璃形成。改进未来模型的一种方法可以是将稀有事件取样技术集成到MD模拟代码中，以大规模地延长平衡时间尺度，并更准确地模拟真实的玻璃系统中的玻璃化和结构配置。
We report new time-resolved synchrotron x-ray diffraction (SXRD) measurements to track structural transformations in calcium-aluminate (CaO)(x)(Al2O3)(1?x) liquids during glass formation, and review recent progress in neutron diffraction with isotope substitution (NDIS) experiments, combined with aspherical ion model molecular dynamics (AIM-MD) simulations, to identify the atomic-scale configurational constraints on glass-forming ability. The time-resolved measurements reveal substantial changes in ordering on short- and intermediate-range occurring during supercooling. In the equimolar composition x??=??0.5 (CA), the liquid undergoes a remarkable structural re-organisation on vitrification as over coordinated AlO5 polyhedra and oxygen triclusters breakdown to form a network of predominantly corner-shared AlO4 tetrahedra. This is accompanied by the formation of branched chains of edge-and face-sharing Ca-centred CaOy? polyhedra contributing to cationic ordering on intermediate length-scales. The Ca-rich end-member of the glass-forming system x??=??0.75 (C3A) is largely composed of AlO4 tetrahedra, but % unconnected AlO4 monomers and Al2O7 dimers are present, representing a threshold after which the glass can no longer support the formation of an infinitely connected network. Overall, the AIM-MD simulations are in excellent agreement with the SXRD and NDIS experiments suggesting an accurate potential model. However, small discrepancies between the simulated glass structures and experimental measurements are apparent, indicating a small degree of liquid-like ordering persists in the simulated glass trajectories. This may be due to the short simulation time-scales which are unrepresentative of the viscous kinetic processes involved in supercooling and glass formation. One approach to improve future models could be the integration of rare event sampling techniques into MD simulation codes to massively extend equilibration time-scales and more accurately model vitrification and structural configurations in real glass systems.