Stochastic density functional theory: Real- and energy-space fragmentation for noise reduction

Stochastic density functional theory (sDFT) is becoming a valuable tool for studying ground-state properties of extended materials. The computational complexity describing the Kohn–Sham orbitals replaced by introducing set random (stochastic) leading to linear and often sub-linear scaling certain observables at account statistical error. Schemes reduce noise are essential, example, determining structure using forces obtained from sDFT. Recently, we have introduced two embedding schemes mitigate fluctuations in electron resultant on nuclei. Both techniques were based fragmenting system either real space or slicing occupied into energy windows, allowing significant reduction fluctuations. For chemical accuracy, further required, which could be achieved increasing number stochastic orbitals. However, convergence relatively slow as error scales 1 / N χ according central limit theorem, where Nχ In this paper, combined mentioned above new approach that builds overlapped fragments windows. significantly lowers properties, such density, total energy, nuclei, demonstrated G-center bulk silicon.