Halo8: a dataset of chemical reaction pathways incorporating halogen chemistry

# Halo8 Database

Halo8 is a ASE-based database containing total 20,116,288 structures, each with properties computed by DFT.

You can directly access the common properties below with the following code:

Coordinates
Energy
Force

```python
from ase.db import connect

db_path = 'your/path/to/db'

with connect(db_path) as db:
    for row in db.select():
        atoms = row.toatoms()    
        coords = atoms.get_positions()        # Coordinates
        energy = atoms.get_potential_energy() # Energy
        forces = atoms.get_forces()           # Force
```

All of the extra properties below are stored in row.data and can be accessed by their keys:

dand_id
HOMO_idx
HOMO_level
LUMO_idx
LUMO_level
Mulliken_charges
Lowdin_charges
Dipole_moment
Nuclear_repulsion_energy
Electronic_energy
One_electron_energy
Two_electron_energy
Exchange_energy
Correlation_energy
Dispersion_correction

```python 
from ase.db import connect

db_path = 'your/path/to/db'

with connect(db_path) as db:
    for row in db.select():
        your_property = row.data['keys_for_property'] 
``` 

The dand_id field follows the format: [Source]_[Formula]_rxn[XXXXX]_[Y]

• Source: T1x (reactions from the Transition1x dataset) or Halogen (newly added halogen-containing reactions)
• Formula: chemical formula of the system
• rxnXXXXX: reaction index number
• Y: structure index within a given reaction

Each reaction pathway consists of 10 images in total: 1 reactant (Y=0), 8 intermediate NEB images, and 1 product (Y=9).

The final block of 8 intermediates plus the fixed endpoints form the converged MEP.

For example, if the Y indices for a reaction are:

0  1   2   3   4   5   6   7   8   9
   10  11  12  13  14  15  16  17   
   18  19  20  21  22  23  24  25   
                 ....               
   178 179 180 181 182 183 184 185  

The converged MEP of this example is: 0 → 178 → 179 → 180 → 181 → 182 → 183 → 184 → 185 → 9 (10 images total)

The initial interpolation path (Y=1-8) and subsequent NEB optimization cycles (Y=10-17, 18-25, ...) are not part of the final pathway. The highest-energy image along the final MEP corresponds to the transition state.

Note that all geometries are optimized at the GFN2-xTB level and should not be interpreted as DFT-level stationary points.