Dependence of the efficiency of spin Hall torque on the transparency of Pt/ferromagnetic layer interfaces

We report that spin current transport across Pt/ferromagnet (FM) interfaces as measured by the spin torques exerted on the FM is strongly dependent on the type and the thickness of the FM layer and on post-deposition processing protocols. By employing both harmonic voltage measurements and spin-torque ferromagnetic resonance measurements, we find that the efficiency of the Pt spin Hall effect in exerting a dampinglike spin torque on the FM corresponds to an effective spin Hall ratio ranging from $<0.05$ to $>0.10$ under different interfacial conditions. The “internal” spin Hall ratio of the Pt thin films used in this study, after taking the interfacial spin transmission factor into account, is estimated to be $\sim 0.30$. This suggests that a careful engineering of $\mathrm{Pt}/\mathrm{FM}$ interfaces can improve the spin Hall torque efficiency of Pt-based spintronic devices. We also note that the dependence on temperature for both vector components of the spin Hall torque is strongly dependent on the details of the $\mathrm{Pt}/\mathrm{FM}$ interface, and that measurements of magnetic damping as a function of FM layer thickness are not generally reliable for determining the true effective spin mixing conductance for the interface.

Figure 1

(a) The magnetic moment per unit area as a function of sputtered Co thickness in $\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}$ samples. (b) Temperature dependence of the magnetic moment per unit area in a $\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}$ sample with 1 nm of Co. (c) The magnetic moment per unit area as a function of sputtered CoFe thickness in both as-deposited (blue squares) and annealed (red triangles) $\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}$ samples. (d) Temperature dependence of the magnetic moment per unit area in an annealed $\mathrm{Pt}/\mathrm{CoFe}/\mathrm{MgO}$ sample with 0.6 nm of CoFe. The effective magnetic anisotropy energy density in terms of $K_{\mathrm{eff}}{t}_{\mathrm{FM}}^{\mathrm{eff}}$ (e) for as-deposited $\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}$ as a function of effective Co thickness and (f) for annealed $\mathrm{Pt}/\mathrm{CoFe}/\mathrm{MgO}$ as a function of effective CoFe thickness. $K_s$ indicates the interfacial anisotropy energy obtained from the intercept of the linear fit (solid lines).

Figure 2

(a) The dampinglike (fieldlike) torque efficiency ${\xi }_{\mathrm{DL}\left(\mathrm{FL}\right)}$ of Pt from as-deposited $\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}$ series (A) samples as a function of Co thickness. Representative error bars due to the device-to-device variation are shown on the data from samples with 1.3 nm of Co. Smaller error bars for other data represent the uncertainties due to fitting. (b) ${\xi }_{\mathrm{DL}\left(\mathrm{FL}\right)}$ of Pt from a representative $\mathrm{Pt}\left(4\right)/\mathrm{Co}\left(1\right)/\mathrm{MgO}\left(2\right)$ series (A) sample as a function of temperature $T$. The blue squares and the red triangles represent data for the dampinglike torque and from fieldlike torque, respectively. For comparison, a dashed line is shown in (b) to indicate the sign and the magnitude of the Oersted field expressed in terms of a spin torque efficiency.

Figure 3

(a) The dampinglike (fieldlike) torque efficiency ${\xi }_{\mathrm{DL}\left(\mathrm{FL}\right)}$ of Pt from $\mathrm{Pt}/\mathrm{CoFe}/\mathrm{MgO}$ series (B) samples annealed at 350 °C as a function of CoFe thickness. The blue squares and the red triangles represent data for the dampinglike torque and the fieldlike torque, respectively. The data for 1 nm of CoFe were obtained from ST-FMR measurements on IPM samples, while the rest were obtained from HR measurements on PM samples. (b) ${\xi }_{\mathrm{DL}\left(\mathrm{FL}\right)}$ of Pt from a representative $\mathrm{Pt}\left(4\right)/\mathrm{CoFe}\left(0.6\right)/\mathrm{MgO}\left(2\right)$ series (B) sample as a function of temperature $T$. For comparison, a dashed line is shown to indicate the sign and the magnitude of the Oersted field in terms of a spin torque efficiency.

Figure 4

(a) The magnetic moment per unit area as a function of CoFe thickness in IPM as-deposited $\mathrm{CoFe}/\mathrm{Pt}$ series (C) (blue squares) and annealed $\mathrm{CoFe}/\mathrm{Pt}$ series (D) (red circles) samples. (b) The effective magnetic anisotropy energy density of as-deposited $\mathrm{CoFe}/\mathrm{Pt}$ series (C) (blue squares) and annealed $\mathrm{CoFe}/\mathrm{Pt}$ series (D) (red circles) samples plotted in terms of $K_{\mathrm{eff}}{t}_{\mathrm{CoFe}}^{\mathrm{eff}}$ as a function of effective CoFe thickness $t_{\mathrm{CoFe}}^{\mathrm{eff}}$. The red open circles represent data from annealed PM $\mathrm{Pt}/\mathrm{CoFe}/\mathrm{MgO}$ series (B) samples. (c) The magnetic moment per unit area as a function of Co thickness in $\mathrm{Ta}/\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}/\mathrm{Ta}$ series (E) (red circles) and $\mathrm{MgO}/\mathrm{Co}/\mathrm{Pt}$ series (F) (blue squares) samples. (d) The effective magnetic anisotropy energy density of $\mathrm{Ta}/\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}/\mathrm{Ta}$ series (E) (red circles) and $\mathrm{MgO}/\mathrm{Co}/\mathrm{Pt}$ series (F) (blue squares) samples plotted in terms of $K_{\mathrm{eff}}{t}_{\mathrm{Co}}$ as a function of effective Co thickness $t_{\mathrm{Co}}^{\mathrm{eff}}$.

Figure 5

(a) The ST-FMR apparent spin torque efficiency ${\xi }_{\mathrm{FMR}}$ of Pt from IPM $\mathrm{CoFe}/\mathrm{Pt}$ series (C) and (D) samples as a function of CoFe thickness. The blue and red dashed lines serve as guides to the eye for the as-deposited and annealed $\mathrm{CoFe}/\mathrm{Pt}$ data, respectively. (b) The inverse of the ST-FMR apparent spin torque efficiency $1/{\xi }_{\mathrm{FMR}}$ as a function of the inverse of the effective CoFe layer thickness $1/t_{\mathrm{CoFe}}^{\mathrm{eff}}$ for the series (C) and (D) samples, with linear fits. (c) ${\xi }_{\mathrm{FMR}}$ of Pt from IPM $\mathrm{Ta}/\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}$ series (E) samples and $\mathrm{MgO}/\mathrm{Co}/\mathrm{Pt}$ series (F) samples as a function of Co thickness. The blue and red dashed lines serve as guides to the eye for the $\mathrm{MgO}/\mathrm{Co}/\mathrm{Pt}$ and $\mathrm{Ta}/\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}$ data, respectively. (d) The inverse of the ST-FMR apparent spin torque efficiency $1/{\xi }_{\mathrm{FMR}}$, for the $\mathrm{Ta}/\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}$ series (E) samples as a function of $1/t_{\mathrm{Co}}^{\mathrm{eff}}$. The solid line represents a linear fit.

Figure 6

(a) and (b) The damping constant $\alpha$ of CoFe from the as-deposited $\mathrm{CoFe}/\mathrm{Pt}$ series (C) samples (blue squares) and annealed $\mathrm{CoFe}/\mathrm{Pt}$ series (D) samples (red circles) as a function of (a) CoFe thickness and (b) the inverse of the effective CoFe thickness $1/t_{\mathrm{CoFe}}^{\mathrm{eff}}$. The solid lines represent linear fits. (c) and (d) The damping constant $\alpha$ of Co from the $\mathrm{Ta}/\mathrm{Pt}/\mathrm{Co}/\mathrm{MgO}$ series (E) samples (red circles) and from the $\mathrm{MgO}/\mathrm{Co}/\mathrm{Pt}$ series (F) samples (blue squares) as a function of (c) Co thickness and (d) the inverse of the effective Co thickness $1/t_{\mathrm{Co}}^{\mathrm{eff}}$. The solid lines represent linear fits to the data in the thick Co regime ($t_{\mathrm{Co}}^{\mathrm{eff}}\ge 3$ nm).