Building a consistent and reproducible database for adsorption evaluation in Covalent-Organic Frameworks

We present a workflow that traces the path from the bulk structure of a crystalline material to assessing its performance in carbon capture from coal's postcombustion flue gases. This workflow is applied to a database of 324 covalent−organic frameworks (COFs) reported in the literature, to characterize their CO2 adsorption properties using the following steps:
(1) optimization of the crystal structure (atomic positions and unit cell) using density functional theory,
(2) fitting atomic point charges based on the electron density,
(3) characterizing the pore geometry of the structures before and after optimization,
(4) computing carbon dioxide and nitrogen isotherms using grand canonical Monte Carlo simulations with an empirical interaction potential, and finally,
(5) assessing the CO2 parasitic energy via process modeling.

The full workflow has been encoded in the Automated Interactive Infrastructure and Database for Computational Science (AiiDA). Both the workflow and the automatically generated provenance graph of our calculations are made available on the Materials Cloud, allowing peers to inspect every input parameter and result along the workflow, download structures and files at intermediate stages, and start their research right from where this work has left off. In particular, our set of CURATED (Clean, Uniform, and Refined with Automatic Tracking from Experimental Database) COFs, having optimized geometry and high-quality DFT-derived point charges, are available for further investigations of gas adsorption properties. We plan to update the database as new COFs are being reported.

*** UPDATE December 2019 ***
- Database extended to include 417 COFs (from papers published until September 1st 2019)
- Migration to AiiDA-v1.0.0
- Using the publicly available plugin aiida-lsmo

*** UPDATE February 2020 ***
- Database extended to include 505 COFs (from papers published until February 1st 2020)
- Including AiiDA Groups for quick interactive visualization

*** UPDATE June 2020 ***
- Database extended to include 574 COFs (from papers published until June 1st 2020)

*** UPDATE September 2020 ***
- Include other applications than CCS, considering the same set of 574 COFs

*** UPDATE October 2020 ***
- Database extended to include 626 COFs (from papers published until October 1st 2020)

*** UPDATE February 2021 ***
- Database extended to include 632 COFs (from papers published until February 1st 2021). We added a total of 61 new structures but we discarded 55: see CURATED-COFs GitHub for more details.

*** UPDATE June 2021 ***
- Database extended to include 648 valid COFs (from papers published until June 1st 2021) plus 80 discarded: see CURATED-COFs GitHub for more details.

*** UPDATE July 2022 ***
- Database extended to include 871 valid COFs (from papers published until July 1st 2022) plus 110 discarded: see CURATED-COFs GitHub for more details.

本研究提出一套完整工作流，可覆盖从晶体材料体相结构到其在燃煤后烟道气碳捕获应用中性能评估的全链路。本工作将该工作流应用于文献报道的324个共价有机框架（covalent−organic frameworks, COFs）数据库，通过以下步骤表征其二氧化碳吸附性能：
(1) 采用密度泛函理论（density functional theory, DFT）对晶体结构（原子位置与晶胞）进行优化；
(2) 基于电子密度拟合原子点电荷；
(3) 表征优化前后结构的孔道几何特征；
(4) 采用巨正则系综蒙特卡洛（grand canonical Monte Carlo, GCMC）模拟结合经验相互作用势，计算二氧化碳与氮气的吸附等温线；
(5) 通过过程建模评估二氧化碳捕获的寄生能耗。

完整工作流已在计算科学自动化交互基础设施与数据库（Automated Interactive Infrastructure and Database for Computational Science, AiiDA）中完成编码。本研究的工作流与计算自动生成的溯源图均已在Materials Cloud平台公开，便于同行检视工作流中每一项输入参数与结果、下载中间阶段的结构与文件，并可基于本研究的成果直接开展后续研究。

尤为重要的是，本研究的CURATED（Clean, Uniform, and Refined with Automatic Tracking from Experimental Database，即经实验数据库自动追踪实现清洁、统一与精细化处理）COFs集合已完成几何优化与高质量DFT拟合原子点电荷，可用于后续气体吸附性能的相关研究。本团队将随新报道的COFs不断更新该数据库。

*** 2019年12月更新 ***
- 数据库规模扩展至417个COFs（涵盖截至2019年9月1日发表的文献）
- 完成向AiiDA-v1.0.0版本的迁移
- 采用公开插件aiida-lsmo

*** 2020年2月更新 ***
- 数据库规模扩展至505个COFs（涵盖截至2020年2月1日发表的文献）
- 新增AiiDA分组功能，支持快速交互式可视化

*** 2020年6月更新 ***
- 数据库规模扩展至574个COFs（涵盖截至2020年6月1日发表的文献）

*** 2020年9月更新 ***
- 基于574个COFs的同一数据集，新增碳捕获之外的其他应用场景研究

*** 2020年10月更新 ***
- 数据库规模扩展至626个COFs（涵盖截至2020年10月1日发表的文献）

*** 2021年2月更新 ***
- 数据库规模扩展至632个COFs（涵盖截至2021年2月1日发表的文献）。本研究新增61个新结构，同时剔除了55个不合格结构，详细信息可查阅CURATED-COFs的GitHub仓库。

*** 2021年6月更新 ***
- 数据库规模扩展至648个合格COFs（涵盖截至2021年6月1日发表的文献），另有80个结构被剔除，详细信息可查阅CURATED-COFs的GitHub仓库。

*** 2022年7月更新 ***
- 数据库规模扩展至871个合格COFs（涵盖截至2022年7月1日发表的文献），另有110个结构被剔除，详细信息可查阅CURATED-COFs的GitHub仓库。

搜集汇总

背景与挑战

背景概述

该数据集是一个专注于共价有机框架（COFs）吸附评估的、一致且可重复的数据库，截至2023年11月包含871个有效COFs结构，并提供了优化几何结构和高品质DFT点电荷。数据集通过AiiDA工作流实现全流程可追溯性，支持碳捕获等应用评估，并定期更新以纳入新结构，确保数据的时效性和可靠性。