Machine learning corrected alchemical perturbation density functional theory for catalysis applications
Machine learning corrected alchemical perturbation density functional theory for catalysis applications
复制标题
DOI：
10.1002/aic.17041
复制
发表时间：
2020-10
期刊：
影响因子：
3.7
通讯作者：
Charles D Griego;Lingyan Zhao;K. Saravanan;J. Keith
中科院分区：
文献类型：
--
作者：
Charles D Griego;Lingyan Zhao;K. Saravanan;J. Keith
炼金术微扰密度泛函理论（APDFT）有望实现假想催化剂位点的计算筛选。在此，我们分析了基于Pt（111）参考体系的一阶APDFT计算方案中CHx、NHx、OHx和OOH吸附在不同覆盖范围内的结合能的误差。然后，我们训练了三种不同的支持向量回归机器学习模型，这些模型可以纠正基于碳、氮和氧的三类吸附剂的系统APDFT预测误差。而未经校正的一阶APDFT单独近似准确的吸附质结合能在多达36个假设的合金基于一个3 × 3单位细胞的Pt（111）的单个kon - sham DFT计算，机器学习校正的APDFT将这个数字扩展到超过20,000，并为开发其他基于机器学习的APDFT模型提供了一个配方。
Alchemical perturbation density functional theory (APDFT) has promise for enabling computational screening of hypothetical catalyst sites. Here, we analyze errors in first order APDFT calculation schemes for binding energies of CHx, NHx, OHx, and OOH adsorbates over a range of different coverages on hypothetical alloys based on a Pt(111) reference system. We then train three different support vector regression machine learning models that correct systematic APDFT prediction errors for each of the three classes of carbon, nitrogen, and oxygen based adsorbates. While uncorrected first order APDFT alone approximates accurate adsorbate binding energies on up to 36 hypothetical alloys based on a single Kohn–Sham DFT calculation on a 3 × 3 unit cell for Pt(111), the machine learning‐corrected APDFT extends this number to more than 20,000 and provides a recipe for developing other machine learning‐based APDFT models.