Comment on ‘Chemical bonding in phase-change chalcogenides’

bonding in phase-change chalcogenides use several densities-based and localized orbitals-based descriptors to unveil the nature of chemical bond in these systems from their density functional theory wavefunctions. While the adopted tools are appropriate for the envisaged scope, the conclusions on the nature of bonding in chalcogenides appear to result from an incorrect choice of the fragments to which the bonding descriptors are applied. Indeed, the fragments analysed are inconsistent with either the stoichiometry of the crystals they are derived from or the infinite three-dimensional arrangement of bonds in these crystals. Once a consistent choice of fragments is adopted, the chemical bond in the Ge–Te crystal (the most thoroughly discussed phase-change chalcogenide in the previously mentioned paper) is shown to be electron-deficient and partially multicentre in nature, comprising a mixture of 3c–2e and 2c–1e bonds. This result is in striking contrast with the electron-rich and 3c–4e nature of bonding in such a system claimed in Müller et al (J. Phys.: Condens. Matter 36 325706) paper. Electron deficiency also characterizes all chemical bonds in crystals of the (GeTe)1−x(Sb2Te3)x pseudobinary system series.