We deposited Pt(5 nm)/Co(0.8 nm)/MgO(2 nm) heterostructure sample using magnetron sputter and patterned as 10 µm width Hall bar structure by photolithography technique. In Figure 1a, the sample structure, coordinate systems, and Helium ion irradiation area are depicted. θM is the polar angle between magnetization direction and z-axis, θB is the polar angle between the external magnetic field and z-axis, and ϕ is azimuthal angle from x-axis, and here we ignore the angle differences between the external magnetic field and magnetization directions in the azimuthal angle because of the negligible in-plane anisotropy. After Hall bar fabrication process, Helium ion is irradiated in vertical direction with sample plane having acceleration energy of 30 keV, beam current of 5.5 pA, and dose from 0 to 30 ions/nm2 with step of 10 ions/nm2. Irradiation area covers whole Hall cross for avoiding signal mixing errors by signal differences between irradiation and non-irradiation area as shown red dotted rectangle in Figure 1a (more explanation in Supplementary Figure S1).
Measurement of SOT induced magnetization switching includes three sequences: initialization, SOT writing by pulse current, and reading from Hall resistance measurement. Firstly, sample is under strong enough +z-axis (-z-axis) direction external field to saturate the magnetization in up (down) direction as initial state. After initialized, the current pulse is injected with pulse amplitude (Ip) from -25 to 25 mA (25 to -25 mA) with 1 mA step. During the current pulse injection in x-axis direction in-plane magnetic field is applied to ensure deterministic switching. And the Hall resistance (RH) is measured in the middle of each pulse injection with 100 µA magnitude direct current (DC). Here, we should mention that the RH is unit value calculated from measured Hall voltage (VH) dividing with magnitude of reading current. The RH results as a function of Ip with the previous procedure shows typical hysteresis loops, as shown in Figure 1b-1c, and it indicates SOT induced magnetization switching in PMA system. Here, the shown hysteresis loops in Figure 1b are in case for dose amount of 30 ions/nm2. It is well known that higher in-plane field make switching more easily, but we observe that Helium ion irradiation also reduces the switching current. The SOT driven magnetization switching hysteresis loops for various doses with a fixed 2.2 kOe external in-plane field is shown in Figure 1c. In Figure 1d, the switching currents from magnetization switching loops at each dose and external magnetic field are depicted. In Figure 1d, we found two features. Firstly, in small field region (0.7 kOe), rapid increase of switching current appears. We expect that this increase is caused by the nucleation of multi-domain states during SOT induced magnetization switching process (see the Supplementary Figure S2a for more details). Because of the multi-domains under small field, it is hard to compare the switching current in higher field region directly so that we will not pay attention much. The second feature is main finding of this work in larger field regions (≥1.2 kOe). We found that the switching current is reduced by increasing dose amounts. And at same dose amount, the switching current have linear relation with the external in-plane field strength under the sufficiently smaller in-plane field comparing 1st order effective anisotropy field, which is well-known behavior following \({J}_{C}=\frac{2e}{\hslash }\frac{{M}_{s}{t}_{F}}{{\theta }_{SH}}\left(\frac{{H}_{K,eff}}{2}-\frac{{H}_{x}}{\sqrt{2}}\right)\)28. And here, e is charge of electron, ℏ is Planck constant, Ms is saturation magnetization, tF is thickness of FM layer, HK,eff is the 1st order anisotropy field and Hx is in-plane external magnetic field parallel with current. In order to get better insight of the Helium irradiations effect, we show the switching current reduction ratio at each dose compared with the pristine sample (\(\left|\left({I}_{P.crit}-{I}_{P,crit}^{Dose 0}\right)/{I}_{P,crit}^{Dose 0}\right|\times 100 \%\)) as in Figure 1e. Result shows that the switching current reduction has increasing tendency with dose amount for in-plane external field. The exceptional dependence for small field (0.7 kOe) probably ascribe to the formation of multi-domain state as seen in Supplementary Figure S2b. The reduction ratio appears largely at external field of 3.1 kOe about 14.2%, 25.5%, and 30.3% at dose of 10, 20, and 30 ions/nm2, respectively. Here, the possible physical origins of the switching current reduction can be the enhanced θSH of Pt layers, and/or it can be the decrease of the effective anisotropy field of FM layer. We will discuss more details later.
To understand the more details of switching behavior, the effects of Helium ion irradiation on the magnetic anisotropy fields are investigated. We conducted Anomalous Hall effect (AHE) measurement by swapping the external magnetic field in z-axis direction to obtained normalized Hall resistance (RH(Hext)/RH(Hext=0 Oe)) hysteresis loops, because the AHE signal is proportional to the z-component of magnetization. The normalized AHE hysteresis loops in Figure 2a shows strong enough PMA for all samples. And each coercivity is decreasing (~ 34 %) from 271 Oe to 178 Oe as shown in Figure 2b by increasing dose amounts. In order to obtain the 1st and 2nd order anisotropy fields (HK,eff, HK,2) by using generalized Sucksmith-Thompson (GST) method (see Supplementary Figure S3), we measured normalized AHE by applying in-plane field (Hx) along the current direction as seen in Figure 2c. Here, it must be mentioned that the obtained 1st order anisotropy fields are the effective anisotropy including demagnetization effect, not pure anisotropy field in GST method. The HK,eff and HK,2are shown in Figure 2d as a function of dose. By increasing dose amount from 0 to 30 ions/nm2, HK,eff and HK,2decrease 38.2%, and 27.5%, respectively. We speculate that decrease is mainly caused by interface modulation from Helium ion irradiation process, since the surface anisotropy energy is very sensitive on the quality of the interface between HM and FM layers. Although it is hard to classify and/or probe the effect of the structural modulation caused by the Helium ion irradiation, we can claim that he anisotropic field as well as the coercivity field can be reduced by the Helium ion irradiation. The magnitude of HK,2is only less than half (40.1 %) compared with the HK,eff, however, the SOT analysis without consideration of HK,2 may lead incorrect results29.
Not only the anisotropy characteristics, but also an important parameter in SOT induced magnetization reversal is θSH. The harmonic Hall signal analysis is frequently used method for calculating θSH as well as extracting SOT driven effective fields29,30. It is well known that the SOT has two contributions acting on different directions, so called field-like torque (FLT, ΔHFL) in transverse direction and damping-like torque (DLT, ΔHDL) effective field in longitudinal direction, consideration of AHE and PHE resistances are necessary for obtaining correct results. In Figure 3a, the measured Hall resistance loops are shown at ϕ = 10 to 40° with fixed θB = 80° for dose amount of 30 ions/nm2 sample. Since the AHE and PHE contribute to the measured Hall signal as following the equation30,
$${V}_{H}={I}_{0}{R}_{0}=\frac{{I}_{0}{R}_{AHE}}{2}\text{cos}{\theta }_{M}+\frac{{I}_{0}{R}_{PHE}}{2}{\text{sin}}^{2}{\theta }_{M}\text{sin}2\varphi$$
1
The clear asymmetries are observed for the Hall loops in Figure 3a in the large field. The asymmetry also increases because of the larger PHE contribution for large ϕ. And by adding and subtracting divided asymmetric Hall loop between +B to -B part and -B to +B part, one can separate the contributions of AHE and PHE as seen in Figure 3b and Figure 3c, respectively. Details of extracting method for AHE and PHE resistances is explained in Supplementary Figure S4. Here, Figure 3b and Figure 3c show the data at the angle of θB = 80° and ϕ = 40° at each dose amount. The AHE resistance can be calculated using AHE contribution at θB = 0°, corresponding to zero external in-plane field, and the PHE resistance can be also calculated taking linear plot on the PHE contribution from the slope of sin2θM, in Equation (1). From those measurement analyses, the calculated AHE and PHE resistances are shown in Figure 3d together. RAHE increased from 1.09 to 1.20 Ω (9.8 %) with increasing dose amounts, while RPHE varied within the range of 0.37 to 0.34 Ω. Since the ratio of R=RPHE/RAHE has an important role in analysis of the harmonic Hall measurement result, we calculated the ratio and it changes from 0.34 to 0.29 at dose of 0 and 30 ions/nm2, as depicted in Figure 3e. It must be mentioned that the variation of R with Helium ion irradiation is not significant comparing to other physical quantities. One possible explanation is that the bulk magnetic properties are relatively insensitive on the Helium ion irradiation, while the surface properties, such as surface anisotropy, are more sensitive.
To obtain the θSH or SOT induced effective fields, we measure harmonic Hall with alternating current (AC) of 5.5 mA peak amplitude and 401 Hz frequency (IAC = I0 sin2πft). Because the harmonic Hall measurement is influenced by the Joule heating effect caused by current flow, we follow the four-direction method for eliminating some thermoelectric artifacts31. The 1st and 2nd harmonic Hall loop is measured swapping magnetic field with fixed θB = 85° and ϕ = 0° for ΔHDL and ϕ = 90° for ΔHFL measurements. Each Hall loop result is shown in Figure 4a-4c, respectively. Here, the 1st and 2nd harmonics are measured simultaneously with two lock-in amplifiers at each ϕ and dose amounts. Harmonic Hall voltage signal under AC follows the equation,
$${V}_{H}={I}_{o}{R}_{H}={V}^{1\omega }\text{sin}\left(\omega t\right)-{V}^{2\omega }\text{c}\text{o}\text{s}\left(2\omega t\right)$$
2
Although 1st voltage has almost same signal at each ϕ, 2nd voltage has completely different signals as shown Figure 4b and Figure 4c. These results are come from the different contribution between DLT and FLT. The 2nd order harmonic Hall voltage at each ϕ = 0° and 90° with consideration of 2nd order PMA energy follow the expression29,
$${V}_{x}^{1\omega }={V}_{y}^{1\omega }={V}_{AHE}\text{cos}{\theta }_{M}$$
3
$${V}_{x}^{2\omega }=\frac{{V}_{AHE}}{2} \left({A}_{1}{\Delta }{H}_{DL}-{B}_{1}{\Delta }{H}_{FL}\right)$$
4
$${V}_{y}^{2\omega }=\frac{{V}_{AHE}\text{cos}{\theta }_{M}}{2} ({B}_{1}{\Delta }{H}_{DL}-{A}_{1}{\Delta }{H}_{FL})$$
5
$${A}_{1}\equiv \frac{\text{s}\text{i}\text{n}{\theta }_{M}}{{H}_{K,eff}\text{cos}2{\theta }_{M}-{H}_{K,2}\text{sin}{\theta }_{M}\text{sin}3{\theta }_{M}+{H}_{ext}\text{cos}\left({\theta }_{M}-{\theta }_{H}\right)}$$
6
$${B}_{1}\equiv \frac{R\text{s}\text{i}{\text{n}}^{2}{\theta }_{M}}{{H}_{ext}\text{sin}{\theta }_{H}}$$
7
Following the Equation (2) to (7), we can rewrite DLT and FLT effective fields (ΔHDL, ΔHFL) as function of θM from the measured harmonic Hall voltages as shown in Figure 4d and Figure 4e, respectively. Here, θM can be calculated using experimentally obtained the 1st order harmonic Hall signal at each dose with Equation (3). The results show the different dose dependences on ΔHDL and ΔHFL with θM. When the near of θM = 15°, corresponding magnetization angle at external magnetic field of 3.1 kOe in 0 ions/nm2, ΔHDL has small increasing tendency as shown in the inset in Figure 4d but ΔHFL has decreasing tendency at its magnitude according to dose amount. However, when θM > 15°, both effective fields show great increase and complex behavior having a maximum peak point at θM of range from 40° to 45°. According to simple macro-spin SOT model29, there is no magnetization direction dependence on both effective SOT fields. However, there are much experimental evidences of the magnetization direction dependence on the effective SOT fields30,31,32. The higher order term of SOT can be one of the possible origins of complex angular dependence. According to Ref.30, the high order term of SOT is non-negligible and may cause complex angular dependence. In addition, if Helium ion irradiation modulates the higher-order term of the SOT just similar as the higher-order term of the PMA, the change in angular dependence can be estimated as a phenomenon caused by the Helium ion irradiation. Furthermore, another possible approach explaining such magnetization direction dependent effective SOT fields is from the framework of distorted Fermi surface33. The θM dependent effective fields in Figure 4d and Figure 4e are rather complicated angular dependence compared with the theoretical results are based on the free-electron like model Hamiltonian with exchange coupling and Rashba effect. The experimental results reflect realistic band structures so that the more complex angular dependent explanation is acceptable. It is hard to analysis the exact origins separately. However, it is also true that the varying angular dependent effective fields by degree of ion irradiation has been experimentally observed as seen in Figure 4d and Figure 4e. And it is worth to note that if HK,2 is not considered in the calculation, the result has quite different tendency with Figure 4d-4e, suggesting the critical role of the 2nd order anisotropy in precise analysis of harmonic Hall measurement in all range of θM (see the Supplementary Figure S5).
Because θSH is one of the most important material parameters for SOT based devices, understanding the correlation between ion irradiation induced θSH variation and HM layer state is important. In order to reveal the effect of the Helium ion irradiation on HM layer only, we irradiated the Helium ion on single Pt layer with thickness of 5 nm as same conditions introduced in sample fabrication description. We used the 4-probe measurement technique for measuring resistance with temperature range of 5 K to 225 K and calculated resistivity using sample geometry information with measured resistance. Here, the resistivity curve and the method of calculating resistivity at 300 K are explained in Supplementary Figure S6a. The resistivity of Pt (ρPt) at 300 K and 5 K is shown in Figure 5a and we can observe the increasing resistivity according to dose amount, 43.4 to 47.8 µΩ∙cm (109%) in 5 K and 56.8 to 60.9 µΩ∙cm (107%) in 300 K comparing 0 ions/nm2 and 30 ions/nm2. Figure 5b displays the changed ratio of temperature coefficient (αTemp), following ρ = ρ0 (1 + αTemp ∙ (T-T0) at linear resistivity increasing region (T > 50 K), and residual-resistivity ratio (RRR), comparison of resistivities between 300 K and 5 K in here. We can find the decreasing tendency of αTemp and RRR both, it can be interpreted as increased influence of impurity at higher dose. Because the collision time is inversely proportional to the impurity density, decrease of αTemp and RRR value imply that the Helium ion irradiation makes extra scattering sources by structural distortion in Pt layer. The Figure 5c shows the resistivity dependence of θSH, and it can be calculated with ΔHDL at each dose value using following equation34,
$${\theta }_{SH}=\frac{2\text{e}}{\text{\hslash }}\frac{{M}_{s}{t}_{F}{A}_{HM}}{{I}_{0}}\varDelta {H}_{DL}$$
8
Here, AHM is the cross section area of flowing current into HM. We assume that the influence of irradiation on Ms, tF, and AHM is small enough to ignore, because the irradiated dose amount is scarce to cause interlayer deformation19,20,27. So, we calculated the θSH with Ms = 1100 kA m-1, tF = 0.8 nm, AHM = 10 µm × 5 nm and ΔHDL when θM = 15° at each dose amount using Equation (8). Error bar can be calculated as the averaged of values of the front and rear data starting from θM = 15°. It is found that θSH has linear relation with increased resistivity by Helium ion irradiations, as well reported11,35,36. The θSH increases 0.096 to 0.132 with resistivity growth from 56.8 to 60.9 µΩ∙cm in 300 K, about 5 - 6 times greater than resistivity of bulk Pt (10.6 µΩ∙cm in 20 ℃) in literature37. This result suggests that Helium ion irradiation process makes extra scattering sources, and they raise the resistivity of HM layer. And the extra scattering sources cause improvement of θSH, resulting in more effective switching of the magnetization by SOT. Although the energy efficiency in view of operating the device is slightly worse due to the ion irradiation induced resistance increasement, the improved θSH ratio is ~ 4 times greater compared with resistivity increasement ratio. In terms of power (P=RsampleI2) consumption, the change in resistance and SHA has an inverse relationship. As a result, only 87.4%, 59.6%, and 56.0% of power consumption is expected at 10, 20, and 30 ions/nm2, respectively. (See Figure 5d) Here, the Rsample and I are normalized resistance of HM and normalized current by θSH at each dose. Therefore, it means that Helium ion irradiation enables more efficient data writing in terms of energy consumption. And it is worth mentioning that the critical switching current equation shown in ref.28 does not match with our actual experimental value except only linear relationship with in-plane field. We expect because the formula is based on the macro spin model as like well-known Brown paradox38. Furthermore, there are reports that it does not match the actual value in the micron scale sample39,40. That’s why we obtained θSH from the spin-orbit torque effective field measurement (see Figure 5c), not from the switching current density. Nevertheless, it is clear that the Helium ion irradiation leads to a decrease in the HK,eff, an increase in the θSH, and the more efficient the SOT induced magnetization switching.
In summary, we observe that Helium ion irradiation can properly reduce SOT induced switching current in Pt(5)/Co(0.8)/MgO(2) structure. The reduction appears 14.2%, 25.5%, and 30.3% at dose of 10, 20, and 30 ions/nm2 comparing with the pristine sample under the in-plane external magnetic field of 3.1 kOe. For understanding of physical reasons of decreasing tendency of the switching current, we considered two main possible origins of reduction, HK,eff and θSH. From AHE measurement and GST method, we can extract HK,eff and it decreases from 13.7 to 8.5 kOe (38.2%) comparing dose 0 and 30 ions/nm2. Not only HK,eff, θSH also increase from 0.096 to 0.132 (27.2%). Furthermore, it is revealed that improvement of θSH is consequence of increase of Pt resistivity by ion irradiation process. Although the power consumption is slightly worse due to the increase of the resistance, the decreased critical current caused by the improved θSH has a greater impact in power consumption. As a result, the ratio of power requiring for operation of device is calculated to consume only about 56.0% for switching at 30 ions/nm2 compared to pristine sample, and this successful analysis on Helium ion irradiation induced modulation of SOT effect can be expected to improve efficiency of SOT based spintronic devices engineering.