OVERTONE

OVERTONE computes anharmonic vibrational spectra with second-order vibrational perturbation theory (VPT2). It works with any calculator of the Atomic Simulation Environment (ASE), from quantum chemistry programs such as ORCA or PySCF to machine-learning potentials, on-the-fly active learning with FALCON, and variational quantum algorithms on (simulated) quantum computers.

OVERTONE provides fundamentals, overtones and combination bands, IR intensities on request, and plotted spectra including the near-IR.

Authors

Basila Bhanu Pattani Ameerjan, Mads-Peter Verner Christiansen and Wilke Dononelli

Features

• Harmonic analysis with symmetrised Hessian and Eckart projection of translations and rotations
• Quartic force field from energies or forces by finite differences in dimensionless normal coordinates
• VPT2 variants: GVPT2 (default), standard VPT2, deperturbed VPT2 and DCPT2
• Automatic detection of Fermi resonances and variational treatment, including intensity borrowing
• Coriolis contributions, linear molecules, constrained atoms (e.g. adsorbates on surfaces)
• Overtones and combination bands up to any number of quanta
• IR intensities (only computed on request, e.g. not for energy-only ML potentials)
• Lorentzian and Gaussian spectra, log scale for the near-IR, experimental reference overlay
• Hybrid schemes with separate calculators for the Hessian and the anharmonic part
• Runs on a single node or distributed over a queueing system with Shephex

Requirements

• Python 3.11 or later
• NumPy, SciPy, ASE (3.23 or later), pydantic, rich, matplotlib

Optional extras: cluster (shephex), falcon (falcon-md), tutorial (tblite) and test.

Installation

pip install overtone-vpt2

or the latest version from GitHub

pip install git+https://github.com/thequantumchemist/overtone.git

For development, clone the repository and install it in editable mode.

git clone https://github.com/thequantumchemist/overtone.git
cd overtone
pip install -e ".[test]"
pytest

Minimal example

from ase.build import molecule
from ase.optimize import BFGS
from tblite.ase import TBLite
from overtone import calculate_anharmonics

def setup_calculator():
    return TBLite(method="GFN2-xTB", verbosity=0, accuracy=0.01)

atoms = molecule("H2O")
atoms.calc = setup_calculator()
BFGS(atoms).run(fmax=1e-4)

result = calculate_anharmonics(atoms, setup_calculator, "water",
                               derivative_source="forces", intensities=True, max_quanta=3)
result.transitions_table()
result.plot_spectrum(log=True)

Use dry_run=True to see the number of single points before starting a large calculation, and recompute_vpt2(directory, method=...) to repeat only the VPT2 step of a finished run with another method.

Important settings

Finite differences amplify numerical noise of the energies. Converge SCF calculations tightly (e.g. ORCA VeryTightSCF). The docstring of calculate_anharmonics describes every parameter.

Manual

The manual (PDF) describes the installation and every function, class and parameter of OVERTONE with its default value.

Tutorial

The tutorials folder contains six examples, from a one-call workflow with GFN2-xTB to Fermi resonances, hybrid schemes and a comparison with the experimental near-IR spectrum of water. The paper folder contains all calculations of the publication.

License

This project is licensed under the GPL-3.0 License. See the LICENSE file for details.

Citation

Please cite the OVERTONE paper (in preparation) when using the code.