Machine learning-driven preparation of high-performance bio-derived carbon catalyst for conversion of cellulose

本研究分为两个部分。一是建立数据库和材料信息学。在这一部分，我们已经采购了112种原材料，包括植物原料和动物原料。用FTIR(50种原料)、X射线衍射仪(50种原料)和扫描电子显微镜(20种原料)对其进行了表征。将20种原料转化为碳材料。通过对138个商业碳素(2万个数据)的建筑数据和预测结果的验证，确定了描述子的选择和数据的前处理方法。第二部分是催化剂的合成及对纤维素转化的评价。在这一部分中，已经开始了利用合成的催化剂对纤维素转化产物进行评价的准备工作，如反应器的建立和高效液相色谱法的建立。

The research is divided into two parts. One is building database and material informatics. In this part, we have sourced 112 raw materials including both plant-based and animal-based materials. The characterization has been done by FTIR (50 raw materials), XRD (50 raw materials), and SEM (20 raw materials). 20 raw materials were converted into carbon materials. The selection of descriptors and data pre-processing methods have been decided based the demonstration of building data and predictions of 138 commercial carbons (20k data). Part two is the synthesis of catalyst and evaluation for conversion of cellulose. In this part, the preparation for evaluation of products from conversion of cellulose by using the synthesized catalysts have been started, such as setting up the reactor and HPLC.