Slater Transition in the Pyrochlore Cd 2 Os 2 O 7
- PDF / 79,959 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 55 Downloads / 171 Views
Slater Transition in the Pyrochlore Cd2Os2O7 D. Mandrus, J. R. Thompson, and L. M. Woods Solid State Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831; and Department of Physics, The University of Tennessee, Knoxville, TN 37996 ABSTRACT Cd2Os2O7 crystallizes in the pyrochlore structure and undergoes a metal-insulator transition (MIT) near 226 K. Here we present resistivity, heat capacity, and magnetization results on Cd2Os2O7. Both single crystal and polycrystalline material were examined. We also present LAPW electronic structure calculations on Cd2Os2O7. We interpret the results in terms of a Slater transition. In this scenario, the MIT is produced by a doubling of the unit cell due to the establishment of antiferromagnetic order. A Slater transition--unlike a Mott transition--is predicted to be continuous, with a semiconducting energy gap opening much like a BCS gap as the material is cooled below TMIT. INTRODUCTION The synthesis and initial characterization of Cd2Os2O7 were reported in 1974 by Sleight, et al. [1], but no subsequent publications have appeared on this compound. The physical properties reported in Ref. 1 are quite intriguing: Cd2Os2O7 was found to crystallize in the pyrochlore structure and to undergo a continuous, purely electronic metal-insulator transition (MIT) near 225 K. It was also found that the MIT was coincident with a magnetic transition that the authors characterized as antiferromagnetic. The idea that antiferromagnetic ordering can double the unit cell and for a half-filled band produce a metal-insulator transition goes back to Slater, who in 1951 proposed a split-band model of antiferromagnetism [2]. In this model, the exchange field favors up spins on one sublattice and down spins on another sublattice and for large U/W reduces to the atomic model of an antiferromagnetic insulator, with a local moment on each site. The thermodynamics of the metal-insulator transition were worked out in mean-field approximation by Matsubara and Yokota (1954) [3], and by Des Cloizeaux (1959) [4]. In both treatments a continuous metalinsulator transition was predicted, with a semiconducting gap opening much the way a BCS gap opens in a superconductor. These results have, of course, been largely supplanted by modern SDW theory [5], but it must be remembered that SDW theory is unambiguously effective only in the weak coupling limit. For example, at half filling it should be possible to proceed smoothly from a weak coupling SDW state to a Mott insulating state as U/W is increased. Although strong coupling SDW theory gets some aspects of the Mott insulating state right, it is wrong in other respects and must be regarded as an incomplete description [5]. One of the major failings of strong coupling SDW theory is that the gap is predicted to disappear above the Neel temperature, whereas in true Mott insulators like CoO the gap persists despite the loss of long range magnetic order. In such materials a Mott-Hubbard description is clearly correct, but when the metal-insulator transition temperature a
Data Loading...
