mofscreen 1.0.2

Automated DFT screening of MOFs for Li-ion anode material properties using CP2K

mofscreen

Automated DFT screening of Metal-Organic Frameworks (MOFs) for Li-ion anode material properties using CP2K.

Calculates four key properties from a single CIF file:

#	Property	Method
1	Electronic bandgap	Single-point DFT
2	Li adsorption energy	GEO_OPT (MOF + Li)
3	Formation energy	Instant (reuses #1)
4	Volume expansion	Instant (reuses #2)

---

Prerequisites

This library requires CP2K to be installed and accessible on your system.

# Install CP2K via conda (recommended)
conda create -n dft_env python=3.11
conda activate dft_env
conda install -c conda-forge cp2k ase numpy
pip install mofscreen

---

Installation

pip install mofscreen

---

Quick Start — Python API

from mofscreen import MOFScreener

screener = MOFScreener(
    cif_path      = "my_mof.cif",         # your relaxed CIF file
    cores         = 16,                    # CPU cores to use
    cp2k_data_dir = "/home/user/miniconda/envs/dft_env/share/cp2k/data",
)

# ── Run everything (recommended) ──────────────────────────────
results = screener.run_all()

print(f"Bandgap       : {results.bandgap.bandgap_ev:.3f} eV")
print(f"Classification: {results.bandgap.classification}")
print(f"E_ads (Li)    : {results.adsorption.e_ads_ev:.4f} eV")
print(f"E_form/atom   : {results.formation.e_form_per_atom_ev:.4f} eV/atom")
print(f"Volume exp.   : {results.volume.expansion_pct:.2f} %")

---

Run Individual Calculations

You can run any single property without running the full pipeline:

from mofscreen import MOFScreener

screener = MOFScreener(
    cif_path      = "my_mof.cif",
    cores         = 16,
    cp2k_data_dir = "/path/to/cp2k/data",
)

# ── Bandgap only ───────────────────────────────────────────────
bg = screener.calc_bandgap()
print(f"Gap: {bg.bandgap_ev:.3f} eV  [{bg.classification}]")
print(f"HOMO: {bg.homo_ev:.3f} eV | LUMO: {bg.lumo_ev:.3f} eV")

# ── Li adsorption only (inserts 2 Li ions) ─────────────────────
ads = screener.calc_adsorption(n_li=2)
print(f"E_ads: {ads.e_ads_ev:.4f} eV")

# ── Formation energy only ──────────────────────────────────────
fm = screener.calc_formation()
print(f"E_form/atom: {fm.e_form_per_atom_ev:.4f} eV/atom")

# ── Volume expansion only ──────────────────────────────────────
vol = screener.calc_volume()
print(f"Expansion: {vol.expansion_pct:.2f} %")

---

Advanced Options

screener = MOFScreener(
    cif_path      = "my_mof.cif",
    cores         = 32,
    mpi_ranks     = 4,              # hybrid MPI + OpenMP
    cp2k_data_dir = "/path/to/data",
    high_accuracy = True,           # TZV2P basis (publication quality)
    fast_mode     = False,          # set True for quick screening
)

results = screener.run_all(
    n_li         = 4,      # insert 4 Li ions
    cell_opt     = True,   # relax cell vectors (true volume expansion)
    compute_refs = True,   # compute self-consistent elemental references
)

---

Command-Line Interface

After installation, mofscreen is available as a CLI command:

# Full pipeline
mofscreen my_mof.cif --cores 16 --cp2k-data ~/miniconda/envs/dft_env/share/cp2k/data

# Bandgap only
mofscreen my_mof.cif --cores 16 --cp2k-data /path/to/data --only bandgap

# Adsorption with 4 Li ions
mofscreen my_mof.cif --cores 16 --cp2k-data /path/to/data --only adsorption --n-li 4

# High accuracy + compute references
mofscreen my_mof.cif --cores 16 --cp2k-data /path/to/data --high-accuracy --compute-refs

# Set data dir via environment variable instead
export CP2K_DATA_DIR=/home/user/miniconda/envs/dft_env/share/cp2k/data
mofscreen my_mof.cif --cores 16

All CLI options

Flag	Default	Description
--cores / -n	16	OMP threads per process
--mpi-ranks	1	MPI ranks (multi-node)
--cp2k-data	—	Path to CP2K data directory
--only	—	Run one calc: bandgap, adsorption, formation, volume
--n-li	1	Number of Li ions to insert
--cell-opt	off	Relax cell during adsorption
--high-accuracy	off	TZV2P basis set
--fast	off	Lower cutoffs (400 Ry)
--compute-refs	off	Compute elemental references
--li-ref-ev	auto	Li reference energy (eV/atom)
--ref-energies	—	JSON file with reference energies

---

Finding Your CP2K Data Directory

# After conda install cp2k:
conda activate dft_env
python -c "import subprocess, shutil; p=shutil.which('cp2k'); print(p)"

# Typical locations:
# ~/miniconda/envs/dft_env/share/cp2k/data
# ~/anaconda3/envs/dft_env/share/cp2k/data
# /usr/share/cp2k/data

# Verify it contains the right files:
ls ~/miniconda/envs/dft_env/share/cp2k/data/BASIS_MOLOPT

---

Output Files

All outputs are saved in a results/ folder next to your CIF file:

results/
├── bandgap.inp          # CP2K input for bandgap
├── bandgap.out          # CP2K output for bandgap
├── adsorption.inp       # CP2K input for adsorption
├── adsorption.out       # CP2K output for adsorption
├── mof_with_li.cif      # MOF structure with inserted Li
├── elemental_refs/      # Elemental reference calculations
│   └── ref_energies.json
├── summary.json         # All results in JSON format
└── run.log              # Full log of the run

---

Result Fields Reference

BandgapResult

Field	Type	Description
bandgap_ev	float	Bandgap in eV (PBE — underestimates by ~30-50%)
classification	str	METALLIC, SEMI-METAL, SEMICONDUCTOR, INSULATOR, etc.
homo_ev	float	HOMO energy in eV
lumo_ev	float	LUMO energy in eV
scf_converged	bool	True if SCF converged
total_energy_ev	float	Total DFT energy in eV

AdsorptionResult

Field	Type	Description
e_ads_ev	float	Adsorption energy: E(MOF+Li) − E(MOF) − n×E(Li)
relaxed	bool	True if GEO_OPT converged
n_ions	int	Number of Li ions inserted

FormationResult

Field	Type	Description
e_form_ev	float	Total formation energy in eV
e_form_per_atom_ev	float	Formation energy per atom in eV/atom
refs_complete	bool	True if all elemental references were available

VolumeResult

Field	Type	Description
expansion_pct	float	Volume expansion in % after Li insertion
v_before_A3	float	Volume of bare MOF in ų
v_after_A3	float	Volume with Li in ų

---

License

MIT License

---

Citation

If you use this library in your research, please cite it appropriately.