Supporting Information for the Journal Article "High-throughput ab initio reaction mechanism exploration in the cloud with automated multi-reference treatment"

This dataset contains the supporting information for the journal article "High-throughput ab initio reaction mechanism exploration in the cloud with automated multi-reference treatment". It consists of two files:

• "structures.json", containing 3335 unique structures found in the exploration, and
• "energies.json", containing 2227 elementary steps and its energies found in the exploration

In the file "structures.json", the following keys are used to characterize a structure:

•  _id: the ID of the structure
• nAtoms: the total number of atoms in the structure
• atoms: the XYZ coordinates of the atoms in atomic units
• charge: the total charge of the structure
• multiplicity: the spin multiplicity 2S+1 of the structure
• label: the position of the structure on the minimum energy pathway

In the file "energies.json", there are two major fields: structures and reactions.

In the field "structures", the keys are:

• id: the ID of structure (can be linked to XYZ coordinates in "structures.json")
• label: the position of the structure on the minimum energy pathway
• energy_dft: PBE-D3BJ/def2-SVP total energy in atomic units
• energy_hf: HF/cc-pVDZ total energy in atomic units
• energy_ccsd: CCSD/cc-pVDZ total energy in atomic units
• energy_ccsd_t: CCSD(T)/cc-pVDZ total energy in atomic units
• energy_gibbs_correction: PBE-D3BJ/def2-SVP Gibbs free energy correction in atomic units

In the field "reactions", the keys are:

• id: the ID of the elementary step
• lhs: the structure IDs on the left-hand side of the reaction
• rhs: the structure IDs on the right-hand side of the reaction
• type: regular or barrierless
• barrierless: binary key to indicate whether the elementary step has a barrier
• ts: ID of the transition state structure
• reaction: the ID of the reaction to which the elementary step belongs

Due to the nature of exploration, certain structures can be explored from multiple elementary steps. Therefore, duplicated structures are common in the reaction exploration. In "structures.json", we de-duplicated the structures and only list the unique ones. The coordinates (in bohr) and global attributes like spin multiplicity and charge are provided. 

In "energies.json", each "step" represents an elementary step. 1260 of the steps are barrierless, while 967 of the steps have a barrier. The structures keys cited in "lhs", "rhs", and "ts" are before de-duplication. The "id" in each structure key points to the unique, de-duplicated structure in "structures.json".

The main catalytical pathways (figure 9) can be found in the dataset with the following elementary_step IDs:

• 62066d2fb8e290f3afb607e2
• 6206704db8e290f3afb607e3
• 62067db7b8e290f3afb607e4
• 620a6d49b8e290f3afb607e5
• 620a6dbeb8e290f3afb607e6
• 620a6e2fb8e290f3afb607e7
• 620a6e78b8e290f3afb607e8
• 620a7f01b8e290f3afb607e9
• 620a7f54b8e290f3afb607ea