Dehybridization transition in intermetallic transition-metal compounds

Publisher Website
Google Scholar
Abstract
The classical s–d theory of Mott for transition metals is reconsidered. The s and d states hybridize at low temperatures, and the electron–phonon coupling constant λ of the hybridized state is dominated by the d component. As the temperature rises, the electron–phonon scattering rage of the d states, τdd −1, exceeds the hybridization integral J sd (more precisely, h(τ−1 dd−-τss −1) > 2 J sd), and as a result the s and d components of the wavefunction become dehybridized, forming decoupled s and d channels, as in the original Mott theory. This process is described using a simple Drude-like theory, which turns out to be somewhat analogous to motional narrowing in NMR and EPR. In specific transition-metal intermetallic compounds, the value of the hybridization integral J sd, derived from the electronic band structure, is small (10–50meV), and as a result the dehybridization takes place at rather low temperatures (100–200K), accounting for anomalies in the resistivity observed there experimentally. At higher temperatures the scattering rate of the s electrons is given by τ−1 sd + τss −1 = J 2 sdτdd/h2 + 2πλb T/h where λs, is the electron-phonon coupling of the s channel, and τdd saturates a value h/ΔE d, the inverse of the d bandwidth. This model applies to intermetallic compounds possessing the A-15 structure, to valence-fluctuation compounds, possibly to materials considered in the past to be spin-fluctuation compounds, to Chevrel phases, and in general to many intermetallic compounds with transition-metal elements.