Random Forest Predictor for Diblock Copolymer Phase Behavior
Random Forest Predictor for Diblock Copolymer Phase Behavior
复制标题
DOI：
10.1021/acsmacrolett.1c00521
复制
发表时间：
2021-10-14
影响因子：
7.015
通讯作者：
Olsen, Bradley D.
中科院分区：
文献类型：
--
作者：
Arora, Akash;Lin, Tzyy-Shyang;Olsen, Bradley D.
基于物理的模型是模拟嵌段共聚物相行为的主要方法。然而，成功地使用自洽场理论（SCFT）设计新材料依赖于正确的化学和温度相关的Flory-Huggins相互作用参数chi（AB），它量化了两个块A和B之间的不相容性，以及精确估计库恩长度（B（A）/B（B））和块密度的比值。这项工作使用机器学习通过直接使用嵌段的化学身份来模拟AB二嵌段共聚物的相行为，避免了测量chi（AB）和B（A）/B（B）的需要。在对4768个数据点的数据集进行训练后，采用的随机森林方法以几乎90%的准确度预测相行为，几乎是使用SCFT采用来自基团贡献理论的chi（AB）获得的准确度的两倍。机器学习模型对测量分子参数的不确定性非常敏感;然而，即使对于高度不确定的实验测量，其准确度仍然保持至少60%。当外推到训练集之外的化学物质时，准确性大大降低。这项工作表明，随机森林相位预测器在许多情况下表现非常好，提供了一个机会来预测自组装分子参数的测量。
Physics-based models are the primary approach for modeling the phase behavior of block copolymers. However, the successful use of self-consistent field theory (SCFT) for designing new materials relies on the correct chemistry- and temperature-dependent Flory-Huggins interaction parameter chi(AB) that quantifies the incompatibility between the two blocks A and B as well as accurate estimation of the ratio of Kuhn lengths (b(A)/b(B)) and block densities. This work uses machine learning to model the phase behavior of AB diblock copolymers by using the chemical identities of blocks directly, obviating the need for measurement of chi(AB) and b(A)/b(B). The random forest approach employed predicts the phase behavior with almost 90% accuracy after training on a data set of 4768 data points, almost twice the accuracy obtained using SCFT employing chi(AB) from group contribution theory. The machine-learning model is notably sensitive toward the uncertainty in measuring molecular parameters; however, its accuracy still remains at least 60% even for highly uncertain experimental measurements. Accuracy is substantially reduced when extrapolating to chemistries outside the training set. This work demonstrates that a random forest phase predictor performs remarkably well in many scenarios, providing an opportunity to predict self-assembly without measurement of molecular parameters.