adcc

Official Resources

• Homepage: https://adc-connect.org/
• Documentation: https://adcc.readthedocs.io/
• Source Repository: https://github.com/adc-connect/adcc
• License: BSD 3-Clause License

Overview

adcc (Algebraic Diagrammatic Construction-connect) is a Python-based hybrid library designed to perform excited-state calculations using the Algebraic Diagrammatic Construction (ADC) scheme for the polarization propagator. It serves as a flexible backend that can interface with various SCF codes (like PySCF and Psi4) to obtain the Hartree-Fock reference, enabling correlated excited-state calculations up to the third order of perturbation theory.

Scientific domain: Correlated electronic structure, excited states, propagator methods, ADC Target user community: Developers and researchers needing modular ADC implementations or reliable excited states beyond TDDFT

Theoretical Methods

• ADC(2): Second-order ADC
• ADC(2)-x: Extended second-order ADC
• ADC(3): Third-order ADC
• CVS-ADC: Core-Valence Separation for core states
• Intermediate State Representation (ISR)
• Dyson equation solvers
• Davidson diagonalization

Capabilities (CRITICAL)

• Excitation energies (Singlet and Triplet)
• Transition moments and oscillator strengths
• One-particle density matrices
• Excited state dipole moments
• Core-excited states (XAS/XES) via CVS
• Interactive usage via Python/Jupyter
• Verification against Q-Chem/MOLCAS results

Sources: Official website, GitHub, J. Chem. Phys. 2020

Key Strengths

Modular Design:

• Host-code agnostic (needs matrix/tensor interface)
• Easy Python integration
• Tensor-based implementation (using opt_einsum)

Methodological Range:

• From fast ADC(2) to accurate ADC(3)
• Rigorous handling of double excitations
• Size-consistent excited states

Core Spectroscopy:

• CVS implementation essential for X-ray absorption
• Target specific edges
• Accurate relaxation effects

Inputs & Outputs

• Input formats:

  • Python objects (SCF results from PySCF/Psi4)
  • Python scripts

• Output data types:

  • Excitation tables (Pandas dataframes)
  • Spectral data
  • Analyzing state character (doubles %, etc.)
  • Densities

Interfaces & Ecosystem

• Host interactions: PySCF (native support), Psi4 (native support), VeloxChem
• Language: Python (frontend), C++ (backend kernels)
• Dependencies: NumPy, H5Py, opt_einsum, pybind11

Advanced Features

Interactive Analysis:

• Analyze state composition interactively
• Plot spectra directly in notebooks
• Export densities for visualization

High-Order Methods:

• ADC(3) implementation available
• Useful for benchmarking TDDFT
• Captures double excitation character

Performance Characteristics

• Speed: Optimized tensor contractions
• Accuracy: ADC(2) similar to CC2, ADC(3) higher
• System size: Moderate (tens to small hundreds of atoms)
• Parallelization: Shared memory (OpenMP/NumPy threads)

Computational Cost

• ADC(2): N^5 scaling (feasible for molecules)
• ADC(3): N^6 scaling (smaller systems)
• Memory: Tensor storage can be significant
• Intermediate: Requires HF ground state

Limitations & Known Constraints

• Reference: Restricted to HF (Canonical orbitals usually)
• Open-shell: UHF support exists but limited
• Solvation: No implicit solvent models in library itself
• Gradients: Analytical gradients not fully exposed/implemented for all methods

Comparison with Other Codes

• vs Q-Chem: adcc is free/open-source, Q-Chem commercial
• vs TURBOMOLE: adcc offers Python flexibility, ADC(3)
• vs EOM-CCSD: ADC(2) cheaper than EOM-CCSD, often sufficient
• Unique strength: Open-source, hackable, Pythonic ADC(3) implementation

Application Areas

• Benchmarking: Validating TDDFT results
• Charge Transfer: Handling CT states correctly
• Core States: X-ray spectroscopy simulation
• Double Excitations: States with significant non-single character

Best Practices

• Basis Set: Requires decent basis (aug-cc-pVDZ/TZ)
• Memory: Ensure sufficient RAM for tensor intermediates
• CVS: Use large core-valence separation energy gap
• Comparison: Compare ADC(2) vs ADC(3) for convergence

Community and Support

• Open-source BSD
• Developed by Heidelberg/Tubingen groups
• Documentation on ReadTheDocs
• Active GitHub

Verification & Sources

Primary sources:

1. Website: https://adc-connect.org/
2. GitHub: https://github.com/adc-connect/adcc
3. M. F. Herbst et al., J. Chem. Phys. 152, 244119 (2020)

Confidence: VERIFIED - Active academic project

Verification status: ✅ VERIFIED

• Official homepage: ACCESSIBLE
• Source code: OPEN (BSD)
• Method: ADC(n) (Scientifically verified)
• Specialized strength: Python library for high-level excited states