PBE (Perdew-Burke-Ernzerhof) is a generalized gradient approximation (GGA) exchange-correlation functional — the most widely used functional in solid-state DFT calculations.
Why PBE Is Popular
No empirical parameters: Constructed entirely from exact constraints on the exchange-correlation energy
Good balance: Provides reasonable accuracy for structures, energies, and elastic properties across diverse chemistries
Computational cost: Much cheaper than hybrid functionals (HSE06) or many-body methods (GW), allowing calculations on thousands of materials
Consistency: All three major databases (MP, AFLOW, JARVIS) use PBE-based workflows, enabling fair comparisons
Known Limitations
Systematically underestimates band gaps by 30-50%
Overbinds d-electron and f-electron systems without Hubbard U corrections
Poorly captures van der Waals interactions (layered materials, molecular crystals)
Self-interaction error can delocalize electrons in strongly correlated systems
Alternatives in MatCraft
PBE+U: Used by Materials Project for transition metal oxides
OptB88vdW: Used by JARVIS for better van der Waals description
MBJ (modified Becke-Johnson): Used by JARVIS for more accurate band gaps
HSE06: Hybrid functional giving better band gaps, but too expensive for high-throughput