Computational chemists have long utilized experimental databases of thermodynamic properties for benchmarking method to determine how well new models calculated measured properties. These databases have also been utilized in fitting force-field and semi-empirical models, and with current trends in machine learning they are becoming increasingly valuable. The existence of highly accurate measured thermodynamic properties of gas phase molecules, in particular those containing elements from hydrogen to neon, are intrinsically linked to the development of methods even today. Unfortunately, as one moves further down the periodic table fewer and fewer of these highly accurate gas-phase thermodynamic properties exist. Upon reaching the first row transition metals for example, the largest compiled database of experimental or high accuracy computational data is around only 100 examples many of which are exotic species and do not reflect the sort of compounds most often utilized in computational or experimental systems. We have developed a new database that increases the amount of high accuracy computational calculations for transition metal complexes by an order of magnitude. By utilizing the NWChem software packages optimally designed parallelization methods, with the correlation consistent composite approach for transition metals (ccCA-TM) developed by Angela Wilson et. al. we have obtained a database that can be utilized in the development of new computational methods including novel density functionals geared toward transition metals and new machine learned models for application in the computational chemistry community. In addition, the wealth of computational information will enable correlation of properties and molecular orbital trends across multiple metal systems. Varied formats for accessing the data collected throughout this process has been geared to consider the needs of both the machine learning and computational chemistry communities are being utilized.

Computational database for first-row transition metals

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.