Unveiling the Hidden Influence of Defects via Experiment and Data Science
Unveiling the Hidden Influence of Defects via Experiment and Data Science
复制标题
通过实验和数据科学揭示缺陷的隐藏影响
DOI：
10.1021/acs.chemmater.3c01817
复制
发表时间：
2023
影响因子：
8.6
通讯作者：
Brgoch, Jakoah
中科院分区：
文献类型：
--
作者：
Sambur, Justin;Brgoch, Jakoah
材料化学房间里的大象是，所有的材料都有固有的缺陷和杂质。缺陷可以极大地影响材料的功能特性，因为它会降低某些样品的性能，提高其他样品的性能，或者偶尔产生全新的功能。一个典型的例子是在金属氧化物光催化剂上存在固有的表面氧空位(缺陷)，如二氧化钛。1−3这些表面氧空位对于有效结合反应物分子和促进电荷转移、提高活性是至关重要的。2、4类似的电荷俘获缺陷也可以应用于发光材料。例如，持久发光的荧光粉需要同样的缺陷来捕获处于激发态的电子，这些电子最终通过热激活缓慢释放。这种与缺陷相关的发射过程导致了在紧急标志和儿童玩具中发现的SrAl2O4：Eu2+，Dy3+的著名的在黑暗中发光的特性。6、7减轻了荧光粉中的这些相同缺陷，使其能够在LED照明和显示应用中应用。8针对和控制特定(点)缺陷的形成和浓度显然是获得所需功能所不可缺少的。不幸的是，除非缺陷是研究项目的重点，否则大多数科学家和工程师都会掩盖它们的影响，因为一个主要原因是它们非常难以识别、表征和理解它们如何对整体水平的样品特性做出贡献。在这篇社论中，我们强调了这一“缺陷挑战”，强调它只能通过由计算和实验学家共同支持的横切研究努力来解决。它将需要在高通量计算、数据科学以及最重要的单粒子水平上的光谱学方面的新技术。进一步采用公平(可查找、可访问、可互操作和可重复使用)原则9并思考如何共享和归档大型数据集，将在利用目前未得到充分利用的超富数据集(在信息和实际成本方面)方面发挥至关重要的作用。正在进行大量的工作来建立新的方法来研究缺陷，希望能够缩小这一差距。计算建模的发展已经开始揭开对缺陷如何影响材料性能的基本理解。例如，密度泛函理论(DFT)提供了对缺陷能级及其在半导体和绝缘材料(如SrAl2O4：Eu2+)中的作用的洞察，以了解本征缺陷是如何使其著名的持久发光。自那以后，人们开发了10种先进的计算方法，以更准确地表示含有点缺陷的材料的局部结构，确保模拟出最低能量的几何形状。11在考虑多个缺陷、缺陷聚集和缺陷动力学方面也取得了类似的进展，12，13尽管分子动力学(MD)捕获缺陷物理的能力仍然有限，但分子动力学中的机器学习力场在捕获较大系统的材料化学方面越来越受欢迎。自动化这一计算过程的努力也在进一步提高研究缺陷对一系列系统的影响的能力，从而产生可用于操纵材料物理性质的新的晶体−化学连接。然而，一个挫折是，这些计算需要计算昂贵的混合泛函来考虑电荷局部化，使得在不同数量的系统上执行这些计算变得困难。尽管存在这一现实，但仍有产生巨大影响的潜力，因为…
The elephant in the room of materials chemistry is that all materials inherently have defects and impurities. Defects can tremendously influence a material’s functional properties by deteriorating the performance in some samples, enhancing performance in others, or occasionally contributing to entirely new functionality. A classic example is the presence of intrinsic surface oxygen vacancies (defects) on metal oxide photocatalysts such as TiO2. 1− 3 These surface oxygen vacancies are vital for effectively binding reactant molecules and facilitating charge transfer, boosting activity. 2, 4 Similar charge trapping defects also enable applications in luminescent materials. 5 Persistent luminescence phosphors, for example, require the same defects to trap electrons in the excited state, which are eventually slowly released through thermal activation. This defect-related emission process gives rise to the famous glowin-the-dark property of SrAl 2O4: Eu2+, Dy3+ found in emergency signs and children’s toys. 6, 7 Mitigating these same defects in phosphors enables their application in LED lighting and display applications. 8 Targeting and controlling specific (point) defect formation and concentrations is clearly indispensable for obtaining desired functionalities. Unfortunately, unless defects are the focus of a research project, most scientists and engineers gloss over their impact for one main reason they are incredibly difficult to identify, characterize, and understand how they contribute to ensemblelevel sample properties. In this editorial, we highlight this “defect challenge”, emphasizing that it can only be solved through a cross-cutting research effort supported by computational and experimentalists alike. It will require new techniques in high-throughput computing, data science, and, most importantly, spectroscopy at the single-particle level. Further adopting FAIR (Findable, Accessible, Interoperable, and Reusable) principles 9 and thinking about how large data sets are shared and archived will play a vital role in making use of ultrarich (in information and actual costs) data sets that are currently being under-utilized. Significant effort is going into establishing new approaches to study defects with the hope of closing this gap. Developments in computational modeling have started to unlock a fundamental understanding of how defects influence a material’s properties. Density functional theory (DFT), for example, has provided insight into the defect energy levels and their role in semiconducting and insulating materials like SrAl2O4: Eu2+ to understand how intrinsic defects enable its famous persistent luminescence. 10 Advanced computational methods have since been developed to more accurately represent the local structure of materials containing point defects, ensuring the lowest energy geometries are modeled. 11 Similar progress has been made in considering multiple defects, defect clustering, and defect dynamics, 12, 13 while machine learning force fields in molecular dynamics are gaining popularity for capturing the materials chemistry of larger systems, although the ability for molecular dynamics (MD) to capture defect physics is still limited. Efforts to automate this computational process are also furthering the ability to study the impact of defects on an array of systems, leading to new crystal− chemical connections that can be used to manipulate a material’s physical properties. One setback, however, is that these calculations require computationally expensive hybrid functionals to account for charge localization, making it intractable to perform these calculations on a diverse number of systems. Despite this reality, there is still potential for tremendous impact as …
通过深缺陷态发光淬灭：掺 Ce YAG 中通过氧空位的重组途径
影响因子：
8.6
作者：
Christopher Linderälv;D. Åberg;P. Erhart
通讯作者：
P. Erhart