Study Of Voltage-Controlled Perpendicular Magnetic Anisotropy In Ta/Feb/Mgo And W/Feb/Mgo Nanowires By The Hall Effect Measurements

Voltage-controlled perpendicular magnetic anisotropy (PMA) is a recently-found effect, which may be used in MRAM and all-metal transistor 1 applications as a magnetization-switching mechanism. Conventionally the strength of the voltage-controlled PMA effect is measured in a MTJ, which electrodes are magnetized perpendicularly each other 2 . The angle between magnetization of electrodes is evaluated from the magneto-resistance at different applied voltages. From this data the strength of the voltage-controlled PMA is estimated. The second method is the measurement of the dependence of the coercive field on a gate voltage $^{3-5}$. The second method is the direct measurement the voltage-controlled PMA effect and it can reveal different interesting features of the effect. However, the magnetization switching is thermo-activated process and the required high-precission measurement of the coercive field is difficult. Therefore, the measurements of the voltage-dependence of the coercive field have been reported only for the cases when the change of the coercive field is substantial. It is the case when the sample temperature is near the Curie temperature 3 or the change of the coercive field is larger than 100 Oe 4 or when the easy axes of the sample is near the transition from in-plane to out-plane direction 5 . We have developed a method of a precise measurement of the coercive field. Main merit of this method is that it is able to detect even very small changes of the coercive field. Therefore, the magnitude of the voltage-controlled PMA effect can be measured in a variety of different samples of different structure and of different magnitude of the voltage-controlled PMA. The coercive field is measured by measuring Hall angle in a pulsed magnetic field. The precise value of the coercive field was evaluated from two sets of the measurements. In the first measurement, the magnetic pulse of a gradually-increased amplitude was applied and the switching field was measured. In the second measurement, the magnetic pulse of constant amplitude was applied and the switching probability was measured. A sufficient number of those two combined measurements and a statistical analysis were used to evaluate the coercive field. Details of the proposed method will be described at conference site. The precession of measured coercive field was better than 1 Oe. We have studied the voltage-controlled PMA effect in Ta(2) /FeB(1.1)/MgO, W(3) /FeB(1.1)/MgO, Ta(2) /FeB(0.5)/W(0.8)/FeB(0.5)/MgO and W(3) / (FeB$/ mathrm {W}) _{n} /$FeB(0.2)/MgO multilayers. The width of fabricated nanowires was 100, 200, 400 and 1000 nm. A pair of 80-nm-wide Hall nanoprobes was connected to the middle of nanowire. A thick 6-nm MgO gate oxide and Ta (1) /Ru(5) gate electrode were used (Fig. 1(a)). All studied samples show the increase of the coercive field under a negative gate voltage and the decrease of the coercive field under a positive gate voltage (See Fig.1(b)). The polarity of the voltage-controlled PMA effect is the same as when it was measured by another method 2 . Additionally to change of the coercive field, there is a small change of the Hall angle. The Hall angle increases at a negative gate voltage and decreases at a positive gate voltage. Figure 2(a) shows coercive field of Ta(2) /FeB(0.5)/W(0.8)/FeB(0.5)/MgO 400-nm-wide nanowire as a function of the gate voltage. In the measurement range, the dependence is linear. The magnitude of the voltage-controlled PMA is 11 Oe per 1 V or 1.8 Oe/(V/nm). All fabricated samples shows the magnitude of the voltage-controlled PMA effect in the range between 5 and 11 Oe / V. Figure 2(b) shows the Hall angle as a function of the gate voltage for the same nanowire. The change is about 5 mdeg per 1 V of gate voltage or 2.5% per 1 V. Additionally, there is a hysteresis loop. The CV-measurements indicate that the hysteresis loop might be due to the deep defects in the thick MgO gate oxide. The defects in the MgO gate oxide might be due to a low annealing temperature of our samples $( sim 200 ^{0}mathrm {C})$. Figure 2(c) shows the magnetization switching probability as a function of an external magnetic field. The switching probability is clearly distinguished at different gate voltages. A feature, which is substantially different for all our studied samples, is the value of the Hall angle. The Hall angle in the Ta (2) /FeB(1.1)/MgO was the largest (1200 mdeg). The Hall angle decreases to 30–50 mdeg, when the number of the W/FeB interfaces increases. It is known 6 that the polarity of the Anomalous Hall effect is opposite for electrons and holes. The measurements of the ordinary Hall effect and anomalous Hall effect is a standard method 6 to determine the type of conductivity in a metal and to estimate the position of the Fermi level. The study the dependence of the magnitude of the voltage-controlled PMA effect on the conductivity type (the Hall angle) will be reported at the conference site.