Sagnac interferometry for high-sensitivity optical measurements of spin-orbit torque

Abstract

We adapt Sagnac interferometry for magneto-optic Kerr effect measurements of spin-orbit-torqueinduced magnetic tilting in thin-film magnetic samples. The high sensitivity of Sagnac interferometry permits for the first time optical quantification of spin-orbit torque from small-angle magnetic tilting of samples with perpendicular magnetic anisotropy (PMA). We find significant disagreement between Sagnac measurements and simultaneously-performed harmonic Hall (HH) measurements of spin-orbit torque on Pt/Co/MgO and Pd/Co/MgO samples with PMA. The Sagnac results for PMA samples are consistent with both HH and Sagnac measurements for the in-plane geometry, so we conclude that the conventional analysis framework for PMA HH measurements is flawed. We suggest that the explanation for this discrepancy is that although magnetic-field induced magnetic tilting in PMA samples can produce a strong planar Hall effect, when tilting is instead generated by spin-orbit torque it produces negligible change in the planar Hall signal. This very surprising result demonstrates an error in the most-popular method for measuring spin-orbit torques in PMA samples, and represents an unsolved puzzle in understanding the planar Hall effect in magnetic thin films.Spin-orbit torques (SOTs) [1, 2] are of interest for achieving high-efficiency manipulation of magnetization in magnetic memory technologies. SOTs are produced when a charge current is applied through a channel with strong spin-orbit coupling and generates a transverse spin current; this spin current can exert a spin-transfer torque on an adjacent ferromagnet (FM), allowing for low-power, electrical control of FM order. Memory cells with perpendicular magnetic anisotropy (PMA) are often preferred over their easy-plane counterparts because they may be fabricated at a higher density and are more resilient to stray magnetic fields or device heating. Accurate quantification of SOTs in PMA systems is therefore important for the development of future technologies.Several techniques are commonly used to quantify SOTs in PMA heterostructures [2-9], yet these methods often exhibit significant quantitative discrepancies with one another. The most-commonly-used method for PMA samples, the harmonic Hall (HH) technique, measures the strength of spin-orbit torques by using second-harmonic Hall signals to detect current-induced magnetic deflections relative to the out-of-plane orientation [3, 5,7]. This method is attractive for its simplicity and has been employed in hundreds of published papers, but it sometimes produces discrepancies and even clearly-unphysical torque values when applied to samples with relatively strong planar Hall effects [10][11][12][13][14][15]. Members of our research group have recently suggested that calculating SOTs from PMA HH measurements by ignoring the expected signal from the planar Hall effect provides results for the SOTs in better agreement with HH measurements on samples with in-plane anisotropy [15].Here, we test the influence of the planar Hall effect on HH meas...

Sagnac interferometry for high-sensitivity optical measurements of spin-orbit torque

“…This is expected as long as the Co layer is sufficiently thick for full absorption of the transverse component of the incoming spin current and qualitatively consistent with previous electrical measurements (34). The values of ξ DL obtained by the Sagnac measurements on PMA and in-plane samples are consistent, which is often not the case for electrically based second harmonic Hall measurements of SOT (35). This value that we find for the damping-like SOT efficiency is also in quantitative agreement with ST-FMR measurements with similar Pt resistivity (22,36,37).…”

Deep learning for spin-orbit torque characterizations with a projected vector field magnet

“…The resultant |𝜉 #$ | and |𝜉 A$ | of the W(5)/Co-Fe-B(1.4) sample are found to be ~ 0.38 and 0.056, respectively. Note that a recent work has proposed that PHE induced by SOT gives negligible contribution to second harmonic even when RPHE is sizable compared to RAHE [31]. Under the circumstance, the DL-SOT and FL-SOT effective field can be simply quantified with the following equation [19]: In general, the DL-SOT efficiency obtained from harmonic approach is larger than that obtained from hysteresis loop shift measurement.…”

“…To estimate 𝑅 gUY , 𝜑 angle scans with various small fields are performed with the aid of the trained models. A small angle approximation of 𝑅 gUY is adopted in accordance with the following equation [31]:…”

A Perspective on Electrical Generation of Spin Current for Magnetic Random Access Memories

“…Since the spin currents we consider produce a spin-torque within a few atomic spacings of an interface, the spin current's action on a free layer may further be affected by interface-related inhomogeneous spinexcitation processes. The materials physics origins of many such measurement inconsistencies are being actively investigated [90,[95][96][97].…”

“…This is an accessible and convenient measurement. However, the second harmonic signal contains many components, including that of an interface Rashba-field-like term, an often uncontrolled Joule heating, as well as possibly from details of interface magnetic moment's anisotropy potential [52,90]. These together add complexity to the quantitative determination of the charge-to-spin conversion efficiency.…”