DFTB

Density Functional Tight Binding (DFTB) is an approximate density functional theory method derived from a Taylor series expansion of the Kohn-Sham DFT total energy with respect to charge density fluctuations. The method was originally de…

1. GROUND-STATE DFT 1.5 Tight-Binding VERIFIED 1 paper
Back to Directory Official Website

Overview

Density Functional Tight Binding (DFTB) is an approximate density functional theory method derived from a Taylor series expansion of the Kohn-Sham DFT total energy with respect to charge density fluctuations. The method was originally developed by Elstner, Porezag, Jungnickel, Elsner, Haugk, Frauenheim, Suhai, and Seifert, with the foundational self-consistent-charge (SCC) DFTB method published in Physical Review B in 1998. DFTB provides a computationally efficient alternative to full DFT while

Reference Papers (1)

Full Documentation

Official Resources

  • Homepage: https://dftb.org/ (DFTB+ implementation)
  • Documentation: https://dftbplus.org/documentation
  • License: DFTB method: open; DFTB+ code: LGPL v3

Overview

Density Functional Tight Binding (DFTB) is an approximate density functional theory method derived from a Taylor series expansion of the Kohn-Sham DFT total energy with respect to charge density fluctuations. The method was originally developed by Elstner, Porezag, Jungnickel, Elsner, Haugk, Frauenheim, Suhai, and Seifert, with the foundational self-consistent-charge (SCC) DFTB method published in Physical Review B in 1998. DFTB provides a computationally efficient alternative to full DFT while maintaining much of the physical accuracy.

DFTB consists of a hierarchy of models: DFTB1 (zeroth-order, non-self-consistent), DFTB2 (second-order, self-consistent charge), and DFTB3 (third-order with on-site charge fluctuations). In DFTB1, densities and potentials are written as superpositions of atomic densities, with Kohn-Sham orbitals expanded in localized atom-centered functions. The Hamiltonian and overlap matrices contain only one- and two-center contributions, pre-tabulated as functions of interatomic distance. Higher orders add self-consistent charge redistribution and Coulomb interactions between charge fluctuations without additional adjustable parameters.

Scientific domain: Computational chemistry, materials science, biochemistry
Target user community: Computational scientists requiring efficient large-scale simulations

Theoretical Methods

  • Taylor series expansion of KS-DFT total energy
  • DFTB1 (non-self-consistent, zeroth order)
  • DFTB2 / SCC-DFTB (self-consistent charge, second order)
  • DFTB3 (third-order with on-site charge fluctuations)
  • LCAO with compressed/optimized atomic orbitals
  • Two-center approximation for Hamiltonian matrix elements
  • Mulliken charge representation for charge fluctuations
  • Hubbard parameters (chemical hardness) for self-consistency

Capabilities (CRITICAL)

  • Orders of magnitude faster than full DFT
  • Pre-tabulated Hamiltonian and overlap matrices
  • Self-consistent charge redistribution (DFTB2/DFTB3)
  • No additional adjustable parameters in SCC formalism
  • Treatment of weak interactions (dispersion corrections)
  • Linear response for excited states
  • Compatible with periodic and molecular systems
  • Force and stress calculations for geometry optimization

Inputs & Outputs

Input formats:

  • DFTB+ input files (for DFTB+ implementation)
  • Slater-Koster parameter files (pre-tabulated)
  • Structure files (GEN format, xyz, etc.)

Output data types:

  • Total energies
  • Kohn-Sham-like eigenvalues
  • Mulliken charges and populations
  • Forces and stresses
  • Band structures (for periodic systems)

Interfaces & Ecosystem

  • DFTB+: Primary production implementation (https://dftbplus.org/)
  • Programming language: DFTB+ in Fortran with Python interface
  • Parameter sets: 3ob, pbc, matsci, organc, trans3d, and others
  • Integrations: Interfaces with XTB, CP2K, and other codes
  • Parallel computing: MPI and OpenMP support in DFTB+

Limitations & Known Constraints

  • Accuracy depends on quality of Slater-Koster parameter sets
  • Parameter sets are system-specific (not universally transferable)
  • Cannot match full DFT accuracy for all properties
  • f-element parameterization is challenging
  • Some properties (e.g., band gaps) systematically underestimated

Performance Characteristics

  • 10-100x faster than typical DFT calculations
  • Linear or near-linear scaling with system size
  • Memory efficient due to sparse matrix structure
  • Parallelizable via MPI and OpenMP in DFTB+

Comparison with Other Codes

  • vs DFTB+: DFTB is the method; DFTB+ is the primary code implementation
  • vs xTB (GFN-xTB): xTB uses extended tight binding with different parameterization philosophy
  • vs full DFT: DFTB sacrifices some accuracy for dramatic speed gains
  • vs PM3/AM1: DFTB is derived from DFT principles rather than empirical fitting

Best Practices

  • Choose parameter sets appropriate for your system (3ob for bio/organic, pbc for solids)
  • Converge SCC iterations carefully for systems with significant charge transfer
  • Validate results against DFT for representative test cases
  • Use DFTB3 for systems with significant on-site charge fluctuations
  • Add dispersion corrections for weakly interacting systems

Verification & Sources

Primary sources:

  1. M. Elstner et al., Phys. Rev. B 58, 7260 (1998) - SCC-DFTB
  2. G. Seifert, J. Phys. Chem. A 111, 5609-5613 (2007)
  3. M. Gaus et al., J. Chem. Theory Comput. 7(4), 931-948 (2011) - DFTB3
  4. DFTB+ website: https://dftbplus.org/

Confidence: VERIFIED - Foundational method with extensive literature and production implementation

Related Tools in 1.5 Tight-Binding