Halide perovskites: efficient & accurate DFTB simulations

Dftb perovskites

The fundamental understanding of the material properties of metal halide perovskites is critical to improving the stability and efficiency of future solar cells devices. Computational simulations have been proven to be a valuable tool to describe the physicochemical properties of metal halide perovskites at the atomistic level. For a thorough understanding of these complex systems, the simultaneous modeling of electrons and ions in relatively large systems is required, which is computationally very demanding with DFT methods. Researchers from Eindhoven University of Technology (TU/e) have benchmarked the suitability of the recent GFN1-xTB density functional tight-binding method for simulating halide perovskites properties.

In a recent study the performance of GFN1-xTB has been assessed for computing structural, energetic, vibrational, and optoelectronic properties of several metal halide perovskites. A comprehensive comparison against experiments and DFT calculations has been made for an extensive set of halide perovskites with different chemical composition (cation, metal, anion) as well as structural phase (cubic, tetragonal, orthorhombic).

The paper shows that the GFN1-xTB method can handle larger systems and longer time scales than standard DFT calculations at a fraction of computational cost while maintaining high accuracy in computing a variety of structural and electronic properties. While DFTB displays high general effectiveness for modeling perovskites, there is still room for improvement, most notably the structural distortion after geometry relaxations of certain structural phases and the wrong electronic description of ions containing complex dynamic bonds. With these caveats, the GFN1-xTB method is an effective tool to simulate perovskite properties. With the tunable nature of the xTB Hamiltonian the authors envision improved performance of DFTB through reparametrization.

Relevant links: DFTB, ParAMS

Back

Stay up to date!

Subscribe to our newsletter (5 times a year) to stay up to date about news, job openings, functionality and events such as webinars and workshops!

No sales. No spam. Occasional notifications. Unsubscribe anytime.