Magnetic collapse in transition metal oxides at high pressure: Implications for the earth

Science

Volumn 275, Issue 5300, 1997, Pages 654-657

  Cohen, Ronald E ^a   Mazin, I I ^a   Isaak, Donald G ^b,c  

^a GEOPHYSICAL LABORATORY   (United States)

^b UNIVERSITY OF CALIFORNIA   (United States)

^c AZUSA PACIFIC UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

METAL OXIDE;

ARTICLE; ASTRONOMY; ATMOSPHERIC PRESSURE; GEOLOGY; MAGNETIC FIELD; PRIORITY JOURNAL;

CORE (OUTER); IRON OXIDE; MAGNETIC COLLAPSE; MANTLE (LOWER); TRANSITION METAL OXIDE;

EID: 0030621135     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.275.5300.654     Document Type: Article

References (36)

1. V. R. Murthy, Science 253, 303 (1991).
2. D. Walker, L. Norby, J. H. Jones, ibid. 262, 1858 (1993).
3. D. G. Isaak, R. E. Cohen, M. J. Mehl, D. J. Singh, Phys. Rev. B 47, 7720 (1993).
4. D. M. Sherman and H. J. F. Jansen, Geophys. Res. Lett. 22, 1001 (1995).
5. P. Dufek, P. Blaha, V. Sliwko, K. Schwarz, Phys. Rev. B 49, 10170 (1994).
6. We are not considering magnetic transitions involving ordering and disordering of spins as in a Néel transition, but rather electronic transitions that change the magnitudes of spin moments.
7. S. Ohnishi, Phys. Earth Planet. Inter. 17, 130 (1978).
8. E. C. Stoner, Proc. R. Soc. London Ser. A 169, 339 (1939).
9. G. L. Krasko, Phys. Rev. B 36, 8565 (1987); O. K. Andersen et al., Physica B+C, 86-88 B+C, 249 (1977).
10. G. L. Krasko, Phys. Rev. B 36, 8565 (1987); O. K. Andersen et al., Physica B+C, 86-88 B+C, 249 (1977).
11. J. P. Perdew and Y. Wang, Phys. Rev. B 45, 13244 (1992).
12. We used density functional theory (DFT) methods, which are first principles in that they require no experimental input other than the nuclear charges and positions. LMTO-atomic sphere approximation (ASA) is a fast method that neglects nonspherical contributions to the potential. The LAPW method is among the most accurate electronic structure methods that assume no shape approximations. We compared LMTO-ASA results for CoO and FeO in the LDA approximation with the full-potential LAPW method and found satisfactory agreement. Equations of state are practically the same. The energy differences between high-and low-spin configurations had an error of tens of millirydbergs compared with LAPW, and LMTO transition pressures were about 20 GPa different from LAPW. We found no sensitivity to the number of k points (reciprocal lattice points used in Brillouin zone integrals), atomic sphere radii, or basis set sizes.
13. L Stixrude, R. E. Cohen, D. J. Singh, Phys. Rev. B 50, 6442 (1994).
14. Collapse in B1 is above the structural transition into the B8 structure at high temperatures, although strained B1 can be maintained metastably to high pressures at room temperature [Y. Fei and H.-k. Mao, Science 266, 1678 (1994)].
15. Other available approaches would not be appropriate for this study. Hartree-Fock methods give excellent results for energetics in transition metal oxides at zero pressure but severely overestimate insulating gaps and are inapplicable for metals or for metal-insulator transitions. The B8 structure of FeO is metallic [X. Li and R. Jeanloz, Geophys. Res. Lett. 21, 2183 (1994)]. More exact methods such as GW [F. Aryasetiawan and O. Gunnarsson, Phys. Rev. Lett. 74, 3221 (1995)] are not tractable for the large number of structures and pressures necessary for this study.
16. Other available approaches would not be appropriate for this study. Hartree-Fock methods give excellent results for energetics in transition metal oxides at zero pressure but severely overestimate insulating gaps and are inapplicable for metals or for metal-insulator transitions. The B8 structure of FeO is metallic [X. Li and R. Jeanloz, Geophys. Res. Lett. 21, 2183 (1994)]. More exact methods such as GW [F. Aryasetiawan and O. Gunnarsson, Phys. Rev. Lett. 74, 3221 (1995)] are not tractable for the large number of structures and pressures necessary for this study.
17. O. Gunnarsson, E. Koch, R. M. Martin, Phys. Rev. B 54, R11026 (1996).
18. 3 [D.D. Sarma et al., Phys, Rev, Lett. 75, 1126, (1995)], because the transition metal hopping integrals are ∼40% larger than in binary oxides and the screened U is ∼40% smaller.
19. P. Dufek, P. Blaha, K. Schwarz, Phys. Rev. B 51, 4122 (1995).
20. M. P. Pasternak et al., Phys. Rev. Lett. 65, 790 (1990).
21. W. E. Jackson, E. Knittle, G. E. Brown Jr., R. Jeanloz, Geophys. Res. Lett. 14, 224 (1987).
22. 3 was marginally unstable with respect to periclase plus stishovite.
23. C. B. Bargeron, M. Avinor, H. G. Drickamer, Inorg. Chem. 10, 1338 (1971); H. G. Drickamer and C. W. Frank, Electronic Transitions and the High Pressure Chemistry and Physics of Solids (Halsted, New York, 1973), pp. 126-129.
24. C. B. Bargeron, M. Avinor, H. G. Drickamer, Inorg. Chem. 10, 1338 (1971); H. G. Drickamer and C. W. Frank, Electronic Transitions and the High Pressure Chemistry and Physics of Solids (Halsted, New York, 1973), pp. 126-129.
25. mag = R In(M + 1 ), where R is the gas constant. For FeO, we estimate a bound of dP/oT < 0.015 GPa/K, which is close to the estimate of 0.01 GPa/K of Ohnishi (7). Thus, we expect an increase of about 29 GPa in transition pressure at 2000 K.
26. T. Lay and C. J. Young, Geophys. Res. Lett. 17, 2001 (1990).
27. E. Knittle and R. Jeanloz, Science 251, 1438 (1991).
28. D. E. Loper and T. Lay, J. Geophys. Res. 100 B4, 6397 (1995).
29. H.E. Newsom and K. W. W. Sims, Science 252, 926 (1991).
30. A. Holzheid and H. Palme, Geochim. Cosmochim. Acta 60, 1181 (1996).
31. R. Jeanloz and A. Rudy, J. Geophys. Res. 92, 11433 (1987).
32. N. Uchida and S. Saito, J. Acoust. Soc. Am. 51, 1602 (1972).
33. E. Huang, K. Jy, S.-C. Yu, J. Geophys. Soc. China 37, 7 (1994).
34. V. I. Anisimov, J. Zaanen, O. K. Andersen, Phys. Rev. B 44, 943 (1991).
35. 2g, from our computations.
36. Supported by NSF grant EAR-9418934. Computations were performed on the Cray J90 at the Geophysical Laboratory. D.G.I, was supported in part by the U.S. Office of Naval Research and NSF grant EAR-9405965. We thank O. K. Andersen, Y. Fei, R. Hazen, R. Hemley, H. Krakauer, C. T. Prewitt, D. J. Singh, and L. Stixrude for helpful discussions.