Ingham Michael, Brady Marcus, Crespo-Otero Rachel
Department of Chemistry, University College London, 20 Gordon Street, London WC1H 0AJ, U. K.
J Chem Theory Comput. 2025 Aug 12;21(15):7576-7592. doi: 10.1021/acs.jctc.5c00539. Epub 2025 Jul 24.
Understanding excited-state processes is essential for designing new functional organic materials. Modeling excited states in organic crystals is challenging due to the need to balance localized and delocalized processes and the competition between intramolecular and intermolecular interactions. Cluster models have proven highly effective for describing weakly interacting organic crystals; however, nonperiodic calculations on periodic systems must account for mechanical and electrostatic coupling to the crystal lattice, particularly in cases of extended coordination where covalent bonds are severed, such as in organic polymers and metal-organic frameworks (MOFs). Point charge embedding is a low-cost method for incorporating long-range electrostatics, enabling the consideration of long-range interactions using Ewald embedding. Small clusters have been effective for modeling excited-state processes in MOFs, yet embedding has rarely been included in such studies. In this work, we examine some of the challenges in describing excited states in covalently connected organic crystals using ONIOM(QM:QM') embedding techniques across systems with increasing coordination: diC-BTBT (an organic molecular crystal), polythiophene (an organic polymer), and two MOFs (QMOF-d29cec2 and MOF-5). We analyze the effects of using different electronic structure methods, including TDHF, TDDFT, ADC(2), and CC(2). One of the main challenges is that embedded cluster models are susceptible to overpolarization near the QM:QM' boundary. To address this, we assess the impact of different charge redistribution schemes (Z- ( = 0, 3), RC, and RCD) and implement them in fromage. Additionally, we compare cluster and periodic models. We find that localized models effectively reproduce excited states in both nonconnected systems (diC-BTBT) and fully connected MOFs, whereas polythiophene remains the most challenging due to band conduction. The accuracy of vertical excitations, oscillator strengths, and simulated spectra is strongly influenced by model size, boundary charges, redistribution schemes, and level of theory. We further analyze the effect of vibrational broadening using the nuclear ensemble approach to predict the absorption and emission spectra of MOF-5. Our results provide a heuristic guide for nonperiodic studies of crystalline excited states, highlighting the remarkable relationship between molecular crystals and MOFs, which will be explored in the future work.
理解激发态过程对于设计新型功能性有机材料至关重要。由于需要平衡局域化和离域化过程以及分子内和分子间相互作用之间的竞争,在有机晶体中对激发态进行建模具有挑战性。团簇模型已被证明在描述弱相互作用有机晶体方面非常有效;然而,对周期性系统进行的非周期性计算必须考虑与晶格的机械和静电耦合,特别是在扩展配位的情况下,例如在有机聚合物和金属有机框架(MOF)中,共价键会被切断。点电荷嵌入是一种纳入长程静电作用的低成本方法,能够使用埃瓦尔德嵌入来考虑长程相互作用。小团簇已被证明在对MOF中的激发态过程进行建模方面是有效的,但此类研究中很少包括嵌入。在这项工作中,我们使用ONIOM(QM:QM')嵌入技术,研究了在具有不同配位的系统(二C - BTBT(一种有机分子晶体)、聚噻吩(一种有机聚合物)和两种MOF(QMOF - d29cec2和MOF - 5))中描述共价连接有机晶体激发态时面临的一些挑战。我们分析了使用不同电子结构方法(包括TDHF、TDDFT、ADC(2)和CC(2))的影响。主要挑战之一是嵌入团簇模型在QM:QM'边界附近容易出现过极化。为了解决这个问题,我们评估了不同电荷重新分布方案(Z - ( = 0, 3)、RC和RCD)的影响,并在fromage中实现它们。此外,我们比较了团簇模型和周期性模型。我们发现局域化模型能够有效地重现非连接系统(二C - BTBT)和完全连接的MOF中的激发态,而聚噻吩由于能带传导仍然是最具挑战性的。垂直激发、振子强度和模拟光谱的准确性受到模型大小、边界电荷、重新分布方案和理论水平的强烈影响。我们使用核系综方法进一步分析了振动展宽对预测MOF - 5吸收和发射光谱的影响。我们的结果为晶体激发态的非周期性研究提供了一个启发式指南,突出了分子晶体和MOF之间显著的关系,这将在未来的工作中进行探索。