Optical conductivity of metal alloys with residual resistivities near or above the Mott-Ioffe-Regel limit

Most interesting examples of violations of the Mott-Ioffe-Regel (MIR) resistivity limit are found in materials with strong electronic correlations that are not well understood by theory. We demonstrate that first principles theory can predict the experimentally observed frequency dependence of the optical conductivity for a novel class of metals where the residual resistivity is near or above the MIR limit, which we define as a ``bad metal.'' The predicted optical conductivity of a NiCoCr alloy is in good agreement with experiment. It is demonstrated that the width of the Drude peak describing the low-frequency part of optical conductivity is comparable to the Fermi energy. The latter, together with a mean free path comparable to the interatomic distance, indicates the absence of well-defined quasiparticles. In contrast to traditional bad metals with strong electron-electron interactions, both the high resistivity and the large width of the Drude peak in these alloys result from strong scattering on disordered atomic potentials that can be understood using modern density functionals.