Supporting Information: Kohn-Sham density encoding rescues coupled cluster theory for strongly correlated molecules

Supporting code and analysis for the manuscript:

Kohn-Sham density encoding rescues coupled cluster theory for strongly correlated molecules
Abdulrahman Y. Zamani, Barbaro Zulueta, Andrew M. Ricciuti, John A. Keith, and Kevin Carter-Fenk
(2025)

---

Overview

This repository contains Jupyter notebooks and Python scripts for analyzing bond dissociation energies (BDEs) of first-row transition metal diatomics using CCSD(T) with various DFT reference orbitals (CCSD(T)@DFA) and standard KS-DFT methods.

Key Features

• CCSD(T)@DFA Analysis: Benchmark of CCSD(T)/CBS with HF, SVWN5, PBE, PW91, R²SCAN, and PBE0 reference orbitals
• KS-DFT Comparison: Performance evaluation of 8 density functionals (SVWN5, PBE, PW91, R²SCAN, B3LYP, PBE0, ωB97X-V, ωB97M-V)
• ph-AFQMC Reference: Comparison with phaseless auxiliary-field quantum Monte Carlo data
• Cr₂ PES Analysis: Detailed potential energy surface study of the challenging Cr₂ dimer

---

Repository Structure

.
├── README.md
├── calc_bde_tm_analysis.ipynb    # Main BDE analysis notebook (49 species)
├── cr2_pes_analysis.ipynb        # Cr₂ potential energy surface analysis
├── generate_tables.py            # Standalone Excel export script
│
├── ref_data/                     # Reference data (included)
│   ├── species_ref_data.csv      # Experimental BDEs, spin-orbit corrections
│   └── ph-afqmc_data.csv         # ph-AFQMC reference values
│
├── cr2_multi_diag/               # Output directory for Cr₂ CSV files
│
├── species/                      # ORCA output files (download from Zenodo)
│   ├── Sc-H/
│   ├── Sc-O/
│   ├── ...
│   └── Zn-Zn/
│
└── cr2_pes/                      # Cr₂ PES scan data (download from Zenodo)
    ├── ccsdt_hf/
    ├── ccsdt_pbe/
    ├── ...
    └── exp.txt

---

Data Availability

The ORCA output files (species/ and cr2_pes/ directories) are available on Zenodo:

https://zenodo.org/records/17958091

Download and extract the archive:

# Download from Zenodo
wget https://zenodo.org/records/17958091/files/data.tar.gz

# Extract
tar -xzf data.tar.gz

---

Requirements

Dependencies

python>=3.9
numpy>=1.20
pandas>=1.3
scipy>=1.7
matplotlib>=3.5
plotly>=5.0
openpyxl>=3.0

Installation

# Clone repository
git clone https://github.com/BLZ11/cc-dft.git
cd cc-dft

# Install dependencies
pip install numpy pandas scipy matplotlib plotly openpyxl

# Download and extract data from Zenodo
wget https://zenodo.org/records/17958091/files/data.tar.gz
tar -xzf data.tar.gz

---

Usage

1. BDE Analysis Notebook (calc_bde_tm_analysis.ipynb)

Main analysis notebook for 49 first-row transition metal diatomics:

• M–H: ScH, TiH, VH, CrH, MnH, FeH, CoH, NiH, CuH, ZnH
• M–O: ScO, TiO, VO, CrO, MnO, FeO, CoO, NiO, CuO, ZnO
• M–Cl: TiCl, VCl, CrCl, MnCl, FeCl, CoCl, NiCl, CuCl, ZnCl
• M–H⁺: ScH⁺, TiH⁺, VH⁺, CrH⁺, MnH⁺, FeH⁺, CoH⁺, NiH⁺, CuH⁺, ZnH⁺
• M–M: Sc₂, Ti₂, V₂, Cr₂, Mn₂, Fe₂, Co₂, Ni₂, Cu₂, Zn₂

Features:

• Interactive Plotly figures for data exploration
• Publication-quality matplotlib figures (PDF/PNG export)
• Error statistics (RMSE, MAE, MAX) by bond type
• Comparison with ph-AFQMC reference data

jupyter notebook calc_bde_tm_analysis.ipynb

Outputs:

• fig_benchmark_bar_mpl_with_qmc.pdf — Bar chart for M–O, M–Cl, M–H (with ph-AFQMC)
• fig_benchmark_bar_mpl_no_qmc.pdf — Bar chart for M–H⁺, M–M
• spark_consolidated.pdf — 5×2 matrix of BDE curves

2. Cr₂ PES Analysis Notebook (cr2_pes_analysis.ipynb)

Detailed analysis of the Cr₂ potential energy surface:

• PES curves for all CCSD(T)@DFA methods
• Comparison with experimental PES (Larsson et al., 2022)
• Comparison with best theoretical estimate (BTE)
• T1 diagnostic analysis

jupyter notebook cr2_pes_analysis.ipynb

Outputs:

• Cr-Cr_pes_main.pdf — Main PES comparison figure
• Cr-Cr_pes_SI.pdf — SI PES comparison figure
• cr2_multi_diag/Cr2_{method}.csv — CSV files with PES data and T1 diagnostics for each method

3. Excel Export Script (generate_tables.py)

Standalone script to generate formatted Excel tables:

python generate_tables.py

Output: bde_benchmark_results.xlsx with sheets:

• CC — CCSD(T)@DFA results by bond type
• KS-DFT — DFT results by bond type
• ph-AFQMC — QMC reference data
• Overall (M-H, M-O, M-Cl) — Summary statistics with QMC
• Overall (M-H+, M-M) — Summary statistics without QMC

---

Computational Details

• Basis Set (Geometry Optimization): def2-TZVP
• Basis Set (CCSD(T) CBS Extrapolation): def2-nZVPP (n = T, Q)
• Relativistic Treatment: X2C Hamiltonian; ZORA was used for the BDE calculation of Mn₂
• Spin-Orbit Corrections: Applied from experimental/theoretical references
• Software: ORCA 6.0

---

Citation

If you use this code or data, please cite:

@article{Zamani2025ms,
  title={Kohn-Sham density encoding rescues coupled cluster theory for strongly correlated molecules},
  author={Zamani, Abdulrahman Y. and Zulueta, Barbaro and Ricciuti, Andrew M. and Keith, John A. and Carter-Fenk, Kevin},
  year={2025}
}

And the data repository:

@dataset{Zamani2025data,
  title={Supporting Data: Kohn-Sham density encoding rescues coupled cluster theory for strongly correlated molecules},
  author={Zamani, Abdulrahman Y. and Zulueta, Barbaro and Ricciuti, Andrew M. and Keith, John A. and Carter-Fenk, Kevin},
  year={2025},
  publisher={Zenodo},
  doi={10.5281/zenodo.17958091}
}

---

License

This project is licensed under the MIT License - see the LICENSE file for details.

---

Contact

For questions or issues, please open a GitHub issue or submit a pull request.