Automated Quantum Chemical Pipeline for Organic Laser Molecules

This repository contains a customized Python-based pipeline for high-throughput quantum chemical calculations, specifically designed for execution on the Niagara HPC cluster.

The workflow automates geometry optimization, vibrational frequency checks, vertical excitation energy calculations, and natural transition orbital (NTO) analysis for organic molecules.

⚠️ Note: This pipeline is tailored for use on the Niagara cluster. Please ensure the following dependencies are properly installed and their paths are specified in each .sh script:

• ORCA v5.0.3
• CREST v2.12
• OpenBabel v3.1.1
• TheoDORE v3.1.1
• Jmol v16.2.1

The pipeline was developed and tested using Python 3.10.2.

---

Step-by-Step Workflow

1. (Optional) Validate Raw SMILES Strings

python filter_smiles.py --raw_smiles smiles_raw/ --pp_smiles smiles_pp

2. Generate Canonicalized SMILES File

python create_gen_smiles.py --smiles smiles_pp --smiles_gen smiles_pp/gen_smiles.csv

3. Generate 3D Coordinates and Perform UFF Optimization

Here, smiles.csv provides the SMILES strings of A-B-A type oligomers for example.

python obabel.py
python run_uff.py

For parallelized preprocessing:

python obabel_parallel.py

4. Conformer Search with CREST

python xyz_distribution.py
python run_crest.py

5. Ground-State Geometry Optimization

python run_geo_opt_b97-3c.py

6. Frequency Calculation (B97-3c)

python run_s0_freq_b97-3c.py

7. Check for Imaginary Frequencies

python get_s0_freq_check.py

If any imaginary frequencies are detected, input files for re-optimization will be automatically generated. The script will print a list of affected subdirectories:

Modified subdirectories: ['2', '13', '24', '31', '39', '49', '50', '58', '63', '68', '73', '74']

Paste this list into the following two scripts:

• run_geo_opt_mod_freq_b97-3c.py
• run_s0_freq_b97-3c.py

Example:

lists = ['2', '13', '24', '31', '39', '49', '50', '58', '63', '68', '73', '74']

Then re-run:

python run_geo_opt_mod_freq_b97-3c.py
python run_s0_freq_b97-3c.py

Repeat the frequency check step until no imaginary frequencies remain.

---

8. Extract HOMO–LUMO Energies

python get_geo_opt_result.py

---

9. TD-DFT Calculation: Franck–Condon Properties

python run_vert_exc.py

10. Extract Key Excitation Parameters

python get_vert_exc_result.py

The results, including oscillator strengths, singlet/triplet energies, and spin–orbit coupling, will be saved in:

VEE_b97-3c.csv

---

11. Natural Transition Orbital (NTO) Analysis

Install Dependencies:

pip install numpy networkx
pip install git+https://github.com/jensengroup/xyz2mol.git

Generate and Visualize NTOs:

python nto_analysis_visualization.py
bash nto_analysis_visualization.sh

The NTO figures will be saved as PNGs in:

opt_b97-3c/vert_exc/nto_mulliken/

---

12. Fragment Orbital Density Decomposition

python get_orb_density.py

Result:

orbital_density_mulliken.csv

---

Notes

Output Example
Example theoretical output for SP001 is available on FigShare.

ORCA Dependency
This project uses the ORCA quantum chemistry software, which is not included in this repository.
To run ORCA-dependent scripts, please download ORCA separately from the official website after agreeing to its academic license terms.
We do not distribute ORCA binaries, source, or installation tools.

Production-level calculations
The production-level DFT calculations reported in this study require high-performance computing (HPC) resources and are not feasible on a standard desktop computer, with run times depending strongly on system size and computational settings.

Installation time
The main quantum chemistry packages (ORCA, CREST, TheoDORE, etc.) were accessed via pre-installed HPC modules, so installation time on a desktop computer is not applicable.

---

Citation

This customized pipeline was developed and used in the study:

"A self-driving lab for discovering tunable and soluble organic lasers"

Hyun Suk Park, Mahdi Mazaheri, Changhyeok Choi, Han Hao, Davide Avagliano, Eric Isbrandt, Buddhika S. B. Karunathilaka, Chihaya Adachi, Alán Aspuru-Guzik