MRAM学习笔记——4.SOT-hall器件的测试

前言

最近因为摸鱼摸得不像话，以及测试hall器件的方案理得不清楚，被小导师连着骂了好几顿。
痛定思痛，多刷几篇SOT-hall器件的测试的文章看看。
下面这两篇挺典型的，对器件测得非常细致。
[Cao, Y., Rushforth, A. W., Sheng, Y., Zheng, H., & Wang, K. (2019). Tuning a Binary Ferromagnet into a Multistate Synapse with Spin–Orbit-Torque-Induced Plasticity. Advanced Functional Materials, 29(25), 1–8.]
[Zhang, S., Luo, S., Xu, N., Zou, Q., Song, M., Yun, J., Luo, Q., Guo, Z., Li, R., Tian, W., Li, X., Zhou, H., Chen, H., Zhang, Y., Yang, X., Jiang, W., Shen, K., Hong, J., Yuan, Z., … You, L. (2019). A Spin–Orbit-Torque Memristive Device. Advanced Electronic Materials, 5(4), 1–7.]

这篇博客就把两篇文章测了什么梳理一下，自己后面做测试的时候要做得更细致一些。

测试方案

器件情况

两篇文献，Cao的重点是要把器件的测试性能往SNN(脉冲神经网络)上去应用，Zhang的重点是对测出的结果进行物理解释，所以测试的思路前者是有的放矢的往一个导向上去测，后者是测得更广更全面以期寻找机理差异。
Cao(左图)的膜层结构是：oxidized Si substrate/Ta(0.5 nm)/Pt(3 nm)/Co(1.3 nm)/ Ta(1.2 nm)/Co(4 nm)/AlOx(2 nm)
Zhang(右图)的膜层结构是：oxidized Si substrate/Ta (10 nm)/CoFeB (1.2 nm)/MgO (1.6 nm)/Ta (20 nm)

具体的逐一来看

Paper1：Cao(2019,Advanced Functional Materials)

这篇他要做SNN的突触单元这个应用我就先不细说了，也没有弄很懂，就说他测了啥吧。

测试1

b) RH–Hz loops with µ0Hz scanning from −11.7 mT to µ0Hz max and then back to −11.7 mT.
扫垂直磁场，Rhall-Hz，从初始-11.7mT的扫到不同的终点（u0Hzmax，图注写错了）再扫回-11.7mT

c) The RH–Jx loop with no external field. A sequence of pulses with scanning magnitude from −16.67 × 10¹⁰ to 16.67 × 10¹⁰ A m⁻² and then back to −16.67 × 10¹⁰ A m⁻² is applied. The Jxth and the Jxf denote the threshold and the finishing current densities for the SOT-induced switching, respectively.

Rhall–Jx loop
Jxth：loop开始翻转的Jx
Jxf：loop翻转完成的Jx

测试2

d) RH as a function of the pulse number for sequences of pulses with constant magnitude Jx.
固定Jx为一个电流值，测Rhall – Jx的pulse number的变化关系；然后换不同的Jx再测。
整个测试的pulse的duration和delay是不变的。
每次换Jx测试前，用一个大的负电流Jx_initial=−16.67 × 10¹⁰ A m⁻² 来进行初始化。

测试3

e) Evolution of the EPSP/IPSP (i.e., RH) by applying a sequence of pulse strings with magnitude of 11.67 × 10¹⁰ and −12.5 × 10¹⁰ A m⁻².
为测试EPSP和IPSP（可理解为测试响应延迟吧），而在正负电流之间反复横跳，加20个正，再加20个负，再20个正…

Before applying each sequence of pulses for (d) and (e), an initialization pulse with a magnitude of −16.67 × 10¹⁰ A m⁻² is applied to the channel.
每次换Jx测试前，用一个大的负电流Jx_initial=−16.67 × 10¹⁰ A m⁻² 来进行初始化。

The duration of every pulse is 10 ms, and every Rhall is measured 2 s after the application of each pulse. For the AHE detection of Rhall in (b–e), a small +x directed d.c. current with fixed magnitude of 100 µA is used as the current source. µ0 is permeability in the free space.

即dration=10ms;delay=2s;read current=100uA.

测试4

b) RH as a function of Jx for the Pt/FM1/ Ta/FM2 device (initialized in the RH ≈ 0 state) when only one current pulse with duration of 1 ms is applied.
Rhall-Jx loop 只不过这里加Jx是每次仅一个duration为1ms的pulse。

Before the application of every single pulse or mixed spike in (b), (d), and (f), the sample is switched to an almost zero mz state with RH ≈ 0. The AHE measurement with a small +x directed d.c. current source of 100 µA is taken more than 2 s after the singe pulse or mixed spike ends. This procedure (set RH ≈ 0 → apply pulse/spike → measure RH → set RH ≈ 0) repeats at least five times for each Jx or Δt.

测试5

Figure 3. SOT switching of Pt/FM1/Ta reference sample.
不同u0Hx水平方向辅助磁场下的Rhall-Jx loop测试。
a) Schematic of the Pt/FM1/Ta reference device that consists of a Ta(0.5 nm)/Pt(3 nm) Hall bar with a 6 µm wide Co(1.3 nm)/Ta(1.2 nm)/AlOx(2 nm) square pillar at the cross. During the measurements, an external assistant magnetic field µ0Hx is applied along the +x-direction.

b) RH–µ0Hz loop with µ0Hx = 0. 无水平场的Rhall-u0Hz loop

c) Pulse current-driven RH–Jx SOT switching loop with µ0Hx = 20 mT.

d) Pulse current-driven RH–Jx SOT switching loops with various µ0Hx from 40 to 240 mT.
Loops are shifted vertically for clarity.
The applied pulse sequences for © and (d) are pulse strings with each duration of 10 ms and scanning magnitude between the maximum ±Jx, and all RH data point are measured 2 s after the application of each pulse.
For the AHE detection of RH in (b–d), a small +x directed d.c. current with fixed magnitude of 100 µA is used as the current source.
即dration=10ms;delay=2s;read current=100uA.
u0Hx从40mT变化到240mT，测不同水平场下的Rhall-Jx loop

e) Evolution of Jxth and Jxf with the increasing µ0Hx.

f) µ0Hx dependence of the current controllable region ΔJx (= Jxth −Jxf, the blue circles) and the current controllable proportion ΔJx/Jxf (the green spots).

g) Maximum anomalous Hall resistance RHmax as a function of µ0Hx.

测试环境

Characterization and Measurement: The Kerr characterization of the magnetization hysteresis and domain patterns was taken using a NanoMoke3 magneto-optical Kerr microscopy.
The anomalous Hall effect measurements were carried out at room temperature with Keithley 2602B as the current source and Keithley 2182 as the nanovoltmeter.
The input spikes for STDP measurement as shown in Figure 2c,e were generated using an Agilent B1500A semiconductor device analyzer with the semiconductor pulse generator unit (SPGU).

Paper2：Zhang(2019,Advanced Electronic Materials)

测试1

b) Current-induced switching under a small constant magnetic field Hx = ±50 Oe. Rhall-I loop
c) RH–I loops with varying scanning magnitude between the maximum ±Imax, under Hx = +50 Oe.

A unipolar current pulse with a duration of 0.5 s was used for the current-induced switching in Figure 1b,c. In particular, in Figure 1c, first, the FM was set into a demagnetization state (RH = 0 Ω) excited by a DC current (60 mA) in the absence of a magnetic field and then the magnitude was increased from the current cycles to maximum (from 27 to 51 mA).

测试2

d) Normalized resistance Rhall as a function of current pulses (0.1 ms pulse width) accompanied by the real-time MOKE images of Hall bars.
高清晰度的MOKE不知道在哪家单位可以拍？？

测试3

Figure 2. Tuning RH using amplitude, polarity, duration, and number of consecutive current pulses.
a) Current modulation applied to the sample cor- responding to a sequence of pulses 50 µs long of alternating polarity and increasing amplitude.
b) Resistance modulation resulting from the sequence of pulses described in (a). Current thresholds to observe resistance changes are found for both polarities (Ith = ±46 mA) and indicated by the yellow shaded areas. The inset in (b) shows a zoomed-in image of the change in resistance around the threshold region.
c–f) Phase diagrams of RH as a function of the pulse width and pulse number for four pulse amplitudes c) I = −70 mA, d) −80 mA, e) −90 mA, and f) +70 mA under an in-plane field Hx = +50 Oe.

In the switching phase-diagram measurements, the pulse width was varied from 50 to 500 µs in increments of 10 µs. Before each measurement, the magnet was initialized to a saturated state M↑ (Figure 2c–e) or M↓ (Figure 2f). The initial intermediate state corresponding to RH = 0 Ω in Figure 2b was also excited by a DC current (60 mA) in the absence of a magnetic field.

Rhall-I的三维相图构建测试：改变Pulse width，Pulse number和I的幅值（-70，-80，-90，+70mA） 三个维度。大量测试。

测试4

Figure 3. Resistance tuning obtained by programming consecutive pulse sequences (duration in 150 µs) under Hx = +50 Oe.
按不同的sequence去加Ix的pulse，duration、read current和delay不变，测Rhall。

a) Current-pulse sequence applied to the sample characterized by sequences of positive pulses of fixed number alternating with sequences of negative pulses of increasing number.

b) AHE resistance responses to the current excitation of (a).

c) Current-pulse sequence applied to the sample characterized by sequences of positive pulses of increasing number alternating with sequences of negative pulses of fixed number.

d) AHE resistance responses to the current exci- tation of ©. The magnetization was initially saturated in the downward state by a reset pulse of −90 mA. The pulse amplitudes are +70 and −70 mA for positive and negative values, respectively.

测试5

Figure 4. Resistance and magnetization switching dynamics.
探索磁矩相图变化的测试。固定一个翻转电流Ix，增加pulse number，in-situ去拍摄MOKE。
观察磁矩翻转的路线图，拟合出磁矩整体翻转速度与电流J的关系
a–d) Dependence of the a,b) resistive state and the c,d) corresponding magnetization dynamics on the cumulative pulse time (t) for a,c) down-to-up and b,d) up-to-down switching.
A small in-plane field of +50Oe is applied.
e) Schematic of the magnetic moment orientation near the DW with right-handed chirality.
The black dashed frame represents the region where AHE resistance is measured.
The tilted DW is equalized to a straight line (green dotted line).
f) DW velocity, v, versus current density J.
g) Scaling plot of log v versus J^−0.25.

µ = 0.25

测试环境

Electrical Measurements:
For the anomalous Hall resistance (RH) measurement, a constant 0.1 mA bias read current (IR) was applied using a DC current source (Keithley Model 6221), and the Hall voltage (VR) was measured using a nanovoltmeter (Keithley Model 2182A) in the usual way.
The same current source was used to apply current pulses (I) for the current-induced switching and the phase-diagram measurements.

Magneto-Optic Kerr Effect Imaging:
A MOKE microscope was used to magnetically image the Hall bars.
To observe contrast, first, the magnet was saturated in the −z or +z direction, and an image was taken. This served as a reference image.
Next, a current pulse with a width of 0.1 ms under an in-plane field was applied, and another image was taken.
Finally, the two images were aligned, and the first reference image was subtracted from the second image to generate the final MOKE image, which was used in this work.

最后

测试真的很重要。
器件bonding、测试设备搭建、labview控制程序设置、测试条件调试、测试的现场分析决策等等，学到很多。
欢迎评论指导~