Application of an inverse-design method to optimizing porphyrins in dye-sensitized solar cells.
Application of an inverse-design method to optimizing porphyrins in dye-sensitized solar cells.
复制标题
DOI：
10.1039/c8cp07722c
复制
发表时间：
2019-03
期刊：
影响因子：
--
通讯作者：
Chencheng Fan;M. Springborg;Yaqing Feng
中科院分区：
文献类型：
--
作者：
Chencheng Fan;M. Springborg;Yaqing Feng
染料敏化太阳能电池（DSSC）在过去的几十年中引起了人们极大的兴趣。然而，确定提供最佳功率转换效率（PCE）的有机分子仍然是一个巨大的挑战。在这里，我们应用我们最近开发的，逆设计方法，这个问题的特殊目的，确定卟啉有前途的高PCE。事实证明，这些计算预测出了15种具有最佳性能的新分子，迄今为止还没有研究过这些分子。这些卟啉衍生物将在不久的将来合成，并随后进行实验测试。我们的反向设计方法，PooMa，是基于提供具有最佳性能的分子系统的建议的策略。PooMa是作为一种需要最少资源的工具开发的，因此，它建立在各种近似方法的基础上。它使用遗传算法来筛选数千个（或更多）分子。对于每个分子，密度泛函紧束缚（DFTB）方法用于计算电子性质。在目前的工作中，五个不同的电子性质被确定，所有这些都与光学性能。随后，定量结构-性质关系（QSPR）模型的构建，可以预测PCE通过这五个电子性质。最后，我们通过更精确的DFT计算对我们的结果进行基准测试，这些计算给出了关于预测的最佳分子的进一步信息。
Dye-sensitized solar cells (DSSCs) have attracted much interest during the past few decades. However, it is still a tremendous challenge to identify organic molecules that give an optimal power conversion efficiency (PCE). Here, we apply our recently developed, inverse-design method for this issue with the special aim of identifying porphyrins with promisingly high PCE. It turns out that the calculations lead to the prediction of 15 new molecules with optimal performances and for which none so far has been studied. These porphyrin derivatives will in the near future be synthesized and subsequently tested experimentally. Our inverse-design approach, PooMa, is based on the strategy of providing suggestions for molecular systems with optimal properties. PooMa has been developed as a tool that requires minimal resources and, therefore, builds on various approximate methods. It uses genetic algorithm to screen thousands (or often more) of molecules. For each molecule, the density-functional tight-binding (DFTB) method is used for calculating the electronic properties. In the present work, five different electronic properties are determined, all of which are related to optical performance. Subsequently, a quantitative structure-property relationship (QSPR) model is constructed that can predict the PCE through those five electronic properties. Finally, we benchmark our results through more accurate DFT calculations that give further information on the predicted optimal molecules.