论文标题
分子极化子嵌入波函数
Wavefunction Embedding for Molecular Polaritons
论文作者
论文摘要
极性化学依赖于强烈的光 - 互动现象来改变光腔内的化学反应速率。要解释和理解这些过程,可靠的理论模型的发展至关重要。虽然计算有效的量子电动力学自洽场(QED-SCF)方法,例如量子电动力学密度功能理论(QEDFT)需要准确的功能,而量子电动力学耦合群(QED-CC)方法则可以在准确性方面进行系统性增长,但成本更大。为了克服这种计算瓶颈,我们在此介绍并开发了基于QED-CC-CED-SCF投影基于基于基于的嵌入方法,该方法继承了两个世界中的所有有利属性,即计算效率和准确性。通过研究一些原型但相关的反应,例如甲基转移反应,质子转移反应以及在复杂环境中的质子化反应来评估嵌入方法的性能。使用新嵌入方法获得的结果与更昂贵的QED-CC结果非常吻合。对这些反应进行的分析表明,强度强的相互作用本质上是非常局部的,并且只能在QED-CC水平上处理一个小区域,以捕获由于空腔引起的重要效果。这项工作为极化量子化学方法的未来发展奠定了基础,它将作为开发其他极化嵌入模型的指南。
Polaritonic chemistry relies on the strong light-matter interaction phenomena for altering the chemical reaction rates inside optical cavities. To explain and to understand these processes, the development of reliable theoretical models is essential. While computationally efficient quantum electrodynamics self-consistent field (QED-SCF) methods, such as quantum electrodynamics density functional theory (QEDFT) needs accurate functionals, quantum electrodynamics coupled cluster (QED-CC) methods provide a systematic increase in accuracy but at much greater cost. To overcome this computational bottleneck, herein we introduce and develop the QED-CC-in-QED-SCF projection-based embedding method that inherits all the favorable properties from the two worlds, computational efficiency and accuracy. The performance of the embedding method is assessed by studying some prototypical but relevant reactions, such as methyl transfer reaction, proton transfer reaction, as well as protonation reaction in a complex environment. The results obtained with the new embedding method are in excellent agreement with more expensive QED-CC results. The analysis performed on these reactions indicate that the strong light-matter interaction is very local in nature and that only a small region should be treated at the QED-CC level for capturing important effects due to cavity. This work sets the stage for future developments of polaritonic quantum chemistry methods and it will serve as a guideline for development of other polaritonic embedding models.