Magnetically Induced Valence Transition in Eu1−x SmXS

S. K. Ghatak,K. H. Bennemann
DOI: https://doi.org/10.1007/978-1-4613-0947-5_108
1987-01-01
Abstract:X-ray absorption by Finkel’shetein et al1 suggests that the valency of Sm in Eu1−x Smx S changes from 2+ to nearly 3+ for small x and remains nearly 2+ for larger values of x. This valence transition2 is not caused by internal pressure sihce both SmS and EuS have same lattice constant (5.97 Å). We propose that this valence transition can be understood considering the shift of conduction band due to exchange interaction.The ‘red’ shift of band states results from long range and short range magnetic order and has its origin in the exchange coupling1$$\text{H}_{\text{ex}}=\sum_j\smallint\text{J}(\vec{r}-\vec{\text{R}_j})\vec{\sigma}.\vec{\text{S}}d^3\text{r}$$between conduction electron spin σ and localized spin σ which comes from f-states. For temperature below the magnetic ordering temperature of the host (e.g. EuS) the conduction band is spin-splitted by long range exchange field due to ordered-f electrons3. This splitting is the cause of valence transition of Tm in Eu1−x TmxSe for 0.15 <x<0.75 below TN where Tm f-level is little below the 5d-conduction band4. But at high T (T > Tc or Tw) this first order shift becomes zero, and the short range order only contributes. The valence transition may occur when this shift of 5d-band is enough to close the gap between Sm f-level and conduction band. As the short-range order is mainly due to spin of f-electron of Eu2+ (8 S0), the valence state of Sm depends on the Eu concentrations.
What problem does this paper attempt to address?