Solvation Energies · foundry-ml
foundry_g4mp2_solvation_v1-2
High-accuracy G4MP2 thermochemical and solvation energy dataset across multiple solvents for machine learning.
What foundry_g4mp2_solvation_v1-2 is
foundry_g4mp2_solvation_v1-2 is a tabular quantum-chemical dataset containing composite G4MP2 calculations and density functional theory (DFT) solvation energies. Developed by Logan Ward, Naveen Dandu, Ben Blaiszik, Badri Narayanan, Rajeev S. Assary, Paul C. Redfern, Ian Foster, and Larry A. Curtiss, it provides benchmark molecular properties for aqueous and organic systems.
Evaluating solvation thermodynamics and molecular formation energies at high levels of electronic structure theory is computationally intensive. By compiling relaxed molecular geometries, B3LYP-calculated electronic properties, G4MP2 reference energies, and solvation values across five common solvents, the dataset supports the development of predictive machine learning models in chemistry and materials science.
How it was made
Molecular candidates are indexed using initial SMILES and InChI identifiers generated from CORINA coordinates. Following molecular relaxation, the dataset captures updated SMILES, InChI strings, and Cartesian XYZ coordinates. Electronic and physical features calculated via DFT (B3LYP) include rotational constants, partial atomic charges, dipole moments, isotropic polarisabilities, electronic spatial extent, heat capacities, zero-point vibrational energies, and frontier orbital parameters (HOMO, LUMO, and band gaps).
Target values are computed using high-level methods. These comprise G4MP2 internal energies, atomisation energies, enthalpies, free energies, standard enthalpies of formation at 298 K, and solvation energies determined across acetone, acetonitrile, dimethyl sulfoxide (DMSO), ethanol, and water.
Key facts
- Data format
- Tabular (Parquet)
- Primary tasks
- Tabular regression and tabular classification
- Publication year
- 2022
- DOI
- 10.18126/jos5-wj65
- Solvents
- Acetone, acetonitrile, dimethyl sulfoxide, ethanol, and water
- Methods
- B3LYP and G4MP2
What you can use it for
- Solvation energy prediction
- Train regression models to infer molecular solvation energies across aqueous and organic solvents directly from molecular structures or DFT-level features.
- G4MP2 thermochemical surrogate modelling
- Construct surrogate machine learning algorithms to map cheaper B3LYP electronic properties to high-accuracy G4MP2 formation enthalpies and atomisation energies.
- Multi-solvent transfer learning
- Evaluate model architectures on how well learned representations of solute–solvent interactions transfer between different solvent environments.
Explore the data
The schema, real rows, full-text search and a SQL filter — read in your browser, nothing downloaded.
Explore
Read the real rows without downloading anything
Reading rows…
LiveRead from foundry-ml/foundry_g4mp2_solvation_v1-2 at the moment you asked. Nothing is cached or stored — every row above came from that request.
How the columns are distributed
Over 130,258 rows of default/train
A
float- min
- 0
- median
- 3.09
- mean
- 9.99
- max
- 620k
B
float- min
- 0.34
- median
- 1.37
- mean
- 1.41
- max
- 437.90
C
float- min
- 0.33
- median
- 1.08
- mean
- 1.13
- max
- 282.95
R2
float- min
- 19.00
- median
- 1146.33
- mean
- 1187.69
- max
- 3374.75
alpha
float- min
- 6.31
- median
- 75.58
- mean
- 75.26
- max
- 196.62
atomic_charges
list- min
- 3
- median
- 18
- mean
- 18.03
- max
- 29
bandgap
float- min
- 0.02
- median
- 0.25
- mean
- 0.25
- max
- 0.62
cv
float- min
- 6.00
- median
- 31.57
- mean
- 31.61
- max
- 46.97
g
float- min
- -714.60
- median
- -416.85
- mean
- -410.92
- max
- -40.50
g4mp2_0k
float- min
- -714.03
- median
- -416.34
- mean
- -410.43
- max
- -40.43
g4mp2_atom
float- min
- -4.16
- median
- -2.80
- mean
- -2.80
- max
- -0.35
g4mp2_energy
float- min
- -714.02
- median
- -416.34
- mean
- -410.42
- max
- -40.42
g4mp2_enthalpy
float- min
- -714.02
- median
- -416.33
- mean
- -410.42
- max
- -40.42
g4mp2_free
float- min
- -714.06
- median
- -416.38
- mean
- -410.46
- max
- -40.45
g4mp2_hf298
float- min
- -335.08
- median
- -10.27
- mean
- -8.54
- max
- 240.53
h
float- min
- -714.56
- median
- -416.80
- mean
- -410.88
- max
- -40.48
homo
float- min
- -0.43
- median
- -0.24
- mean
- -0.24
- max
- -0.10
inchi_0
string_text- min
- 17
- median
- 53
- mean
- 52.93
- max
- 86
inchi_1
string_text- min
- 18
- median
- 66
- mean
- 64.28
- max
- 87
lumo
float- min
- -0.17
- median
- 0.01
- mean
- 0.01
- max
- 0.19
mu
float- min
- 0
- median
- 2.48
- mean
- 2.67
- max
- 29.56
n_atom
int- min
- 3
- median
- 18
- mean
- 18.03
- max
- 29
n_electrons
int- min
- 10
- median
- 66
- mean
- 65.90
- max
- 74
n_heavy_atoms
int- min
- 1
- median
- 9
- mean
- 8.80
- max
- 9
smiles_0
string_text- min
- 1
- median
- 15
- mean
- 15.13
- max
- 28
smiles_1
string_text- min
- 1
- median
- 22
- mean
- 22.97
- max
- 59
sol_acetone
float- min
- -72.94
- median
- -8.59
- mean
- -8.85
- max
- 3.05
sol_acn
float- min
- -74.67
- median
- -8.31
- mean
- -8.58
- max
- 3.16
sol_dmso
float- min
- -73.21
- median
- -7.13
- mean
- -7.41
- max
- 4.20
sol_ethanol
float- min
- -82.92
- median
- -8.78
- mean
- -8.96
- max
- 3.67
sol_water
float- min
- -84.35
- median
- -7.30
- mean
- -7.48
- max
- 4.59
u
float- min
- -714.56
- median
- -416.81
- mean
- -410.88
- max
- -40.48
u0
float- min
- -714.57
- median
- -416.81
- mean
- -410.89
- max
- -40.48
u0_atom
float- min
- -4.16
- median
- -2.80
- mean
- -2.80
- max
- -0.34
xyz
string_text- min
- 95
- median
- 565
- mean
- 563.79
- max
- 915
zpe
float- min
- 0.02
- median
- 0.15
- mean
- 0.15
- max
- 0.27
Strengths and limits
Good at
- Pairs high-accuracy G4MP2 benchmark energies with detailed B3LYP electronic and thermodynamic features.
- Covers solvation energy targets across five distinct organic and aqueous solvents for the same molecular systems.
- Supplies both 1D and 2D chemical strings alongside relaxed 3D molecular Cartesian coordinates.
Watch out for
- Only provides a pre-assigned training split, requiring researchers to generate their own evaluation partitions.
- Limited to the specific organic and aqueous solvent systems evaluated in the original calculations.
Questions
Can I look inside foundry_g4mp2_solvation_v1-2 without downloading it?
Yes. This page reads its schema, real rows and per-column statistics in the browser, with full-text search and a SQL filter. Nothing is downloaded to your machine.
What licence is foundry_g4mp2_solvation_v1-2 under?
Its authors declare cc-by-4.0. Read the terms on the original page before reusing the data.
Which solvents are covered in this dataset?
The dataset includes solvation energy values for water, acetone, acetonitrile, dimethyl sulfoxide (DMSO), and ethanol.
What levels of theory are included?
Descriptors and baseline electronic properties are calculated using B3LYP DFT, while target energies and enthalpies of formation are derived using the G4MP2 composite method.
What structural representations are available for each molecule?
Entries contain initial SMILES and InChI identifiers, relaxed SMILES and InChI strings, CORINA coordinates, and post-relaxation 3D Cartesian XYZ coordinates.
How to cite it
Ward, Logan and Dandu, Naveen and Blaiszik, Ben and Narayanan, Badri and Assary, Rajeev S. and Redfern, Paul C. and Foster, Ian and Curtiss, Larry A. DFT Estimates of Solvation Energy in Multiple Solvents. Materials Data Facility, 2022. https://doi.org/10.18126/jos5-wj65
Written from foundry_g4mp2_solvation_v1-2's own dataset card. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.