Abstract
Recent experiments demonstrating large spin-transfer torques in topological-insulator (TI)-ferromagnetic-metal (FM) bilayers have generated a great deal of excitement due to their potential applications in spintronics. The source of the observed spin-transfer torque, however, remains unclear. This is because the large charge transfer from the FM to the TI layer would prevent the Dirac cone at the interface from being anywhere near the Fermi level to contribute to the observed spin-transfer torque. Moreover, there is still not much understanding of the impact on the Dirac cone at the interface from the metallic bands overlapping in energy and momentum, where strong hybridization could take place. Here, we build a simple microscopic model and perform first-principles-based simulations for such a TI-FM heterostructure, considering the strong hybridization and charge-transfer effects. We find that the original Dirac cone is destroyed by hybridization, as expected. Instead, we find an interface state that we dub a "descendent state" that forms near the Fermi level due to the strong hybridization with the FM states at the same momentum. Such a descendent state carries a sizable weight of the original Dirac interface state, and thus it inherits the localization at the interface and the same Rashba-type spin-momentum locking. We propose that the descendent state may be an important source of the experimentally observed large spin-transfer torque in the TI-FM heterostructure.
Original language | English |
---|---|
Article number | 235433 |
Journal | Physical Review B |
Volume | 96 |
Issue number | 23 |
DOIs | |
State | Published - 21 Dec 2017 |
Bibliographical note
Funding Information:Y.-T.H. was supported by the Cornell Center for Materials Research with funding from the NSF MRSEC program (DMR-1719875) and in part by Laboratory for Physical Sciences and Microsoft. E.-A.K. was supported by US Department of Energy, Office of Basic Energy Sciences, Division of Materials Science and Engineering under Award No. DE-SC0010313. E.-A.K. acknowledges Simons Fellow in Theoretical Physics Award No. 392182 and the hospitality of KITP, supported by Grant No. NSF PHY11- 25915. K.P. was supported by the U.S. National Science Foundation Grant No. DMR-1206354. The computational support was provided by SDSC under DMR060009N and VT ARC computer clusters.
Publisher Copyright:
© 2017 American Physical Society.