Topological surface state manipulation of magnetic damping and surface anisotropy in topological insulator/nonmagnet/CoFe heterostructures

The recent advances of high spin-orbit torque efficiency in ferromagnetic/topological insulator (TI) structures hold great promise for the development of high-performance spintronic devices. However, the roles of spin-momentum-locked Dirac surface states (SSs) and the interfacial magnetic proximity effect (MPE) on dynamic magnetic properties are still under debate. Here, we quantitatively distinguish the manipulation effects of SSs and MPE on magnetic damping and surface anisotropy in TI/nonmagnet/CoFe heterostructures. We found that, in addition to the common spin pumping contribution stemming from SSs, damping enhancement also consists of an obvious MPE contribution to the TI/CoFe material system. Moreover, the increased surface magnetic anisotropy for the CoFe films grown on top of a TI layer is believed to arise mainly from the interfacial atomic intermixing. Our paper sheds light on the effects of SSs and MPE on magnetization dynamics, which offer exciting opportunities for developing TI-based spintronics.

Figure 1

(a) Illustration of the sample layer structure and pump-probe laser beams. (b) The typical time-resolved magneto-optical Kerr effect (TR-MOKE) curves for the samples of ${\mathrm{Bi}}_2{\mathrm{Te}}_3, {\mathrm{Sb}}_2{\mathrm{Te}}_3$, and ${\left(\mathrm{CrBiSb}\right)}_2{\mathrm{Te}}_3\left(6\right)$/CoFe (8) under $H=15\mathrm{kOe}$. The inset displays a complete transient TR-MOKE curve for the sample of ${\mathrm{Bi}}_2{\mathrm{Te}}_3$/CoFe. The field dependences of the fitted (c) precession frequency $f$, (d) decay time τ, and (e) the calculated effective magnetic damping factor ${\alpha }_{\mathrm{eff}}$.

Figure 2

(a) The time-resolved magneto-optical Kerr effect (TR-MOKE) curves for the samples of ${\mathrm{Bi}}_2{\mathrm{Te}}_3\left(6\right)/\mathrm{Ta}\left(t_{\mathrm{Ta}}\right)$/CoFe (8) measured under $H=15\mathrm{kOe}$. (b) The saturated damping factor ${\alpha }_{\mathrm{s}}$ as a function of $t_{\mathrm{Ta}}$ for the samples of TI (6)/$\mathrm{Ta}\left(t_{\mathrm{Ta}}\right)$/CoFe (8).

Figure 3

(a) The typical in-plane magnetic hysteresis loops for ${\mathrm{SiO}}_2$/Ta or $\mathrm{Al}\left(t_{\mathrm{NM}}\right)$/CoFe (20) samples. (b) Magnetic coercivity $H_{\mathrm{c}}$ versus the $\mathrm{Ta}\left(t_{\mathrm{Ta}}\right)$ or $\mathrm{Al}\left(t_{\mathrm{Al}}\right)$ interlayer thicknesses for the samples of ${\mathrm{SiO}}_2$/Ta or Al/CoFe (20) and ${\mathrm{SiO}}_2/{\mathrm{Bi}}_2{\mathrm{Te}}_3\left(20\right)$/Ta or Al/CoFe (20).

Figure 4

(a) The typical time-resolved magneto-optical Kerr effect (TR-MOKE) curves under $H=15\mathrm{kOe}$ for the samples of ${\mathrm{SiO}}_2/{\mathrm{Bi}}_2{\mathrm{Te}}_3\left(20\right)$/Ta or Al/CoFe (20) with different interlayer thicknesses and (b) the corresponding field-dependent precession frequencies. The effective saturation magnetization $4\pi M_{\mathrm{eff}}$ and ${\alpha }_{\mathrm{s}}$ as a function of $t_{\mathrm{Ta}}$ or $t_{\mathrm{Al}}$ for the samples of (c) ${\mathrm{SiO}}_2$/Ta or Al/CoFe (20) and (d) ${\mathrm{SiO}}_2/{\mathrm{Bi}}_2{\mathrm{Te}}_3\left(20\right)$/Ta or Al/CoFe (20). Note that the $4\pi M_{\mathrm{eff}}$ values were derived by fitting the field-dependent frequency curves shown in Fig. 4.

Figure 5

The temperature dependences of $4\pi M_{\mathrm{eff}}$ and ${\alpha }_{\mathrm{s}}$ for the samples of (a) ${\mathrm{SiO}}_2$/CoFe (20), ${\mathrm{SiO}}_2/{\mathrm{Bi}}_2{\mathrm{Te}}_3\left(20\right)$/CoFe (20) and (b) ${\mathrm{SiO}}_2/{\mathrm{Bi}}_2{\mathrm{Te}}_3\left(20\right)$/Al(2)/CoFe (20). The inset in (a) shows the corresponding $M_{\mathrm{s}}$ measured at different temperatures.

Figure 6

The in-plane (IP) and out-of-plane (OP) hysteresis loops for samples of (a) ${\mathrm{SiO}}_2/\mathrm{Al}\left(t_{\mathrm{Al}}\right)$/CoFe (20) and (b) ${\mathrm{SiO}}_2/{\mathrm{Bi}}_2{\mathrm{Te}}_3\left(20\right)/\mathrm{Al}\left(t_{\mathrm{Al}}\right)$/CoFe (20).