Official Resources
- Principal Developer: Andrey Kutepov (Brookhaven National Laboratory)
- Host Group: Comsuite (Rutgers/BNL correlation)
- Source Repository: Not publicly open-sourced; likely internal or available upon request (historical link
github.com/flapwmbpt/flapwmbpt is 404).
- Related Project: Comsuite (FlapwMBPT is often cited as the DFT/GW engine within this ecosystem).
Overview
FlapwMBPT is an advanced electronic structure code developed by Andrey Kutepov (BNL) for over 30 years. It combines the rigorous Full-Potential Linearized Augmented Plane Wave (FLAPW) method with Many-Body Perturbation Theory (MBPT) to provide high-precision calculations of ground and excited states. It allows for consistent treatment of electronic correlations using GW approximation, self-consistent GW (scGW), and vertex corrections.
Scientific domain: Strongly Correlated Materials, Actinides, Spectroscopy, Methodological Development.
Target user community: Specialists in F-electron systems (Plutonium, Americium), GW method developers, and correlated electron researchers.
Theoretical Methods
- Basis Set: FLAPW+LO (Full-Potential Linearized Augmented Plane Wave + Local Orbitals).
- Relativity: Scalar-relativistic and Fully Relativistic (Dirac) implementations.
- Many-Body Theory:
- GW Approximation (G0W0, scGW).
- Quasiparticle Self-Consistent GW (qsGW).
- Vertex Corrections (GW+$\Gamma$).
- Dynamic Mean Field Theory (DMFT) integration (via Comsuite).
- Core Solvers: All-electron treatment (no pseudopotentials).
Capabilities
- Electronic Structure: Band structures, DOS, PDOS.
- Excited States: Quasiparticle energies, spectral functions.
- Materials:
- 3d Transition Metals.
- Lanthanides and Actinides (f-electrons).
- Transition Metal Oxides.
- Thermodynamics: Free energy calculations.
- Parallelization: MPI-based parallel execution.
Key Strengths
- All-Electron Accuracy: Removes pseudopotential errors, critical for f-electrons and core states.
- Self-Consistency: Features one of the few fully self-consistent GW implementations.
- Relativistic Effects: High-accuracy treatment of spin-orbit coupling.
- Vertex Corrections: Advanced implementation beyond standard GW.
Inputs & Outputs
- Inputs: Fortran-namelist style inputs (system definition, basis parameters, mixing schemes).
- Outputs:
- Spectral functions ($A(\omega)$).
- Band structures.
- Self-energies ($\Sigma(\omega)$).
- Thermodynamics data.
Performance Characteristics
- Scaling: $N^3$ to $N^4$ depending on the level of theory (GW is expensive).
- Hardware: HPC-oriented, utilizing MPI for scaling across nodes.
- Optimization: Fortran90 codebase optimized for vector processors.
Limitations & Known Constraints
- Availability: Unlike VASP or QE, it is not widely distributed; access is typically via collaboration or specific academic channels.
- Cost: All-electron GW is extremely computationally demanding compared to pseudopotential DFT.
- Documentation: Sparse public documentation; expertise required for usage.
Comparison with Other Codes
- vs WIEN2k: Both are FLAPW. WIEN2k is the community standard for DFT; FlapwMBPT specializes in GW/MBPT.
- vs exciting:
exciting is also all-electron GW (LAPW), but FlapwMBPT has distinct implementations of self-consistency and vertex corrections favored by the Kutepov/Gabriel Kotliar groups.
- vs VASP/QE: FlapwMBPT is all-electron (more accurate but slower) and focuses on many-body physics rather than high-throughput materials search.
Verification & Sources
Primary sources:
- BNL Profile: Andrey Kutepov Research - Confirms development and ownership.
- Comsuite: Cited as a component code in the Comsuite/COMSCOPE project (Rutgers/BNL).
- Literature: Kutepov, A. L., et al. "Electronic structure of Pu and Am metals by self-consistent relativistic GW method." Physical Review B 85 (2012).
Verification status: ✅ VERIFIED (as Research Code)
- Authenticity: Confirmed existence and active research use.
- Accessibility: ⚠️ RESTRICTED (Public repo is missing/dead; code is likely shared privately).