Diode and Selective Routing Functionalities Controlled by Geometry in Current-Induced Spin–Orbit Torque Driven Magnetic Domain Wall Devices

Abstract

Research on current-induced domain wall (DW) motion in heavy metal/ferromagnet structures is crucial for advancing memory, logic, and computing devices. Here, we demonstrate that adjusting the angle between the DW conduit and the current direction provides an additional degree of control over the current-induced DW motion. A DW conduit with a 45° section relative to the current direction enables asymmetrical DW behavior: for one DW polarity, motion proceeds freely, while for the opposite polarity, motion is impeded or even blocked in the 45° zone, depending on the interfacial Dzyaloshinskii–Moriya interaction strength. This enables the device to function as a DW diode. Leveraging this velocity asymmetry, we designed a Y-shaped DW conduit with one input and two output branches at +45° and −45°, functioning as a DW selector. A DW injected into the junction exits through one branch, while a reverse polarity DW exits through the other, demonstrating selective DW routing.

The electric manipulation of magnetic textures is a key area of research for the development of next-generation memory, logic, and in-memory computing devices.¹⁻⁷ In racetrack memory and logic devices, data is encoded within continuous strips comprising a series of magnetic domain walls (DWs), which are precisely displaced by in-plane electric current pulses.^8,9 A particularly significant focus is on perpendicularly magnetized heavy-metal (HM)/ferromagnetic (FM) metallic multilayers, which exhibit broken inversion symmetry and large spin–orbit coupling (SOC). Within these structures, an in-plane current injected into the HM layer generates both damping-like (DL) and field-like (FL) torques on the magnetization of the FM layer.^10,11 Moreover, the SOC induced Dzyaloshinskii–Moriya interaction (DMI) at the HM/FM interface¹² imposes a Néel DW configuration with in-plane DW magnetization aligned parallel to the current, thereby maximizing the efficiency of the damping-like torque in driving the DWs.¹³⁻¹⁵

In straight strips, pulsed current-induced displacement of DWs with alternating polarities progresses in a synchronized stepwise manner. Additionally, reversing the current direction results in an oppositely directed but symmetrical displacement of the DWs.¹⁴ For both racetrack memories and logic devices, breaking the symmetry of alternating polarity DW displacement is of considerable interest as it introduces additional degrees of control over the current-induced DW motion. Upon the application of relatively long current pulses, the DW will tilt due to the canting of the DW magnetization, affecting the current induced domain wall dynamics.¹⁶ This feature can be used in curved, bent, or Y-shape junctions to create an asymmetry in the current induced motion of alternating polarity DWs.¹⁷⁻¹⁹

In this study, we demonstrate that by separating the DW conduit from the current strip and adjusting the angle between the DW conduit and the current direction, it is possible to control current-induced DW motion for short pulses when no significant DW tilt is expected. By carefully tailoring the shape of the DW conduit and the interfacial DMI, we achieve the functionality of a DW diode: the DW moves in one direction under the influence of the current, while motion in the opposite direction is hindered or even blocked. Additionally, we show that Y-shaped DW conduits function as DW selective routers: DWs with one polarity follow one branch of the junction, while DWs with the opposite polarity follow the other branch.

The current-induced chiral DW motion in HM/FM structures is a dynamic phenomenon, which is dependent on the angle between the direction of the current and of the DW motion.²⁰ To qualitatively understand this behavior and how it can be harnessed for the development of chiral DW devices, we employ a simplified model (elaborated in more detail in Supporting Information, section S1). Let us consider an up/down DW situated in the middle of a perpendicularly magnetized HM/FM strip. Initially, in the absence of any current and for strong enough DMI, the DW magnetization (m) aligns along the DMI effective field (H_DMI), pointing toward the negative Ox axis, as indicated schematically in Figure 1(b)-i. When a charge current (J_c) is applied through the strip, the DL torque τ_DL = −γa_jm × (m × p) will act on the DW magnetization. The torque depends on the charge current density (through a_j, defined in the Supporting Information, section S1), on m, and on the spin-current polarization vector p = e_J × n, which is transverse to both the direction of the charge current density (e_J = J_c/J_c) and the direction perpendicular to the strip n = (0,0,1). The action of τ_DL is to rotate m in-plane at an angle α away from the DMI axis (Figure 1(b)-ii). The associated DMI torque τ_DMI = −γm_DW × H_DMI = γH_DMI(sin α)k will rotate the magnetization out-of-plane, causing the DW to move. In principle, the DW velocity depends on the charge current density and the projection of the DW magnetization on the charge current direction, v ∝ J_Ccos(α – β).^13,21 Therefore, by applying a current J_c at various β angles relative to the strip, it is possible to either minimize or maximize the DW velocity. This is shown in Figure 1(a), which depicts the velocity of an up/down DW calculated by using micromagnetic simulations for an ideal strip (see Supporting Information, section S2). During simulations, periodic boundary conditions were assumed along the Oy axis. This ensures that the DW has no tilt, and it remains always parallel with the Oy axis and perpendicular to the Ox axis. Current pulses of density up to 2 × 10¹²A/m² and 10 ns width are applied at 45°, 0°, and −45° with respect to the Ox axis. A DMI constant of |D| = 1.1 mJ/m² was used. At relatively low currents, because τ_DL is small, the magnetization rotation angle α is close to zero. Consequently, as seen in Figure 1(a), the highest velocity occurs when the current is applied along the Ox axis (β = 0), where the current direction and the DW magnetization are most closely aligned compared to the other angles (β = 45° and β = −45°). In contrast, at higher currents, the magnetization rotation angle α becomes significant. For instance, with current density pulses of 1.6 × 10¹²A/m², the DW magnetization rotation angle α, shown in Figure 1(b)ii–iv, indicates that, at β = 45°, the current direction and the DW magnetization are most closely aligned, resulting in maximum velocity. However, when the current is applied along the Ox axis (β = 0), there is less collinearity between the current direction and the DW magnetization, which results in lower velocity. Furthermore, at β = −45°, the current direction and the DW magnetization are least aligned and the velocity is minimum. At the same current density, due to the symmetry, the behavior reverses for a down/up domain: the velocity is maximized at β = −45° and minimized at β = 45°, while at β = 0, the velocity matches that of the up/down DW (Figure 1(c) and (d)).

Figure 1.

(a) Up/down DW velocity derived from micromagnetic simulations for an ideal strip subjected to current pulses of 10 ns width and amplitudes up to 2 × 10¹²A/m², assuming a DMI constant of 1.1 mJ/m². The pulses are applied at β = 45°, 0°, and −45° with respect to the Ox axis. For pulses above 1 × 10¹²A/m², the DW velocity is maximized at β = 45° and minimized at β = −45°. (b) Schematic representation of the DW magnetization m relative to the charge current direction J_c for the current pulses indicated by the shaded area in (a). In the absence of any current, m aligns along the DMI axis. When current is applied, the DL torque tilts the DW magnetization counterclockwise. At β = 45°, J_c and m are relatively well aligned resulting in maximum DW velocity. At β = 0°, the alignment is reduced and the DW velocity decreases, while at β = −45°, minimal alignment leads to minimum velocity. (c) Down/up DW velocity calculated under the same conditions as in (a). Contrary to (a), for pulses above 1 × 10¹²A/m², the DW velocity is maximized at β = −45° and minimized at β = 45°. (d) Schematic representation of the DW magnetization m relative to the charge current direction J_c for the current pulses indicated by the shaded area in (c). In the absence of any current, m aligns along the DMI axis. When current is applied, the DL torque tilts the DW magnetization clockwise. At β = −45°, J_c and m are relatively well aligned, resulting in maximum DW velocity. At β = 0°, the alignment is reduced and the DW velocity decreases, while at β = 45°, minimal alignment leads to minimum velocity.

Additional micromagnetic simulations (see Supporting Information, section S2) indicate that the geometry-induced velocity asymmetry is not restricted to β = −45° and β = 45°, but it is observable when the current is applied at any angle other than zero relative to the strip axis. Moreover, the simulations reveal that, as the width of the DW conduit decreases, the velocity asymmetry not only persists but also increases as the width is reduced into the nanometer range. This indicates that the observed velocity asymmetry is robust and scalable across various DW conduit geometries.

The asymmetry in DW velocity when the current is applied at 45° and −45° relative to the strip suggests that, in practice, by selecting appropriate material parameters, it is feasible to maximize the DW velocity for one current direction while minimizing it, or even blocking DW motion entirely, for the opposite direction. This behavior is corroborated by micromagnetic simulations on realistic systems²² (see Supporting Information, section S2).

To experimentally demonstrate the geometry-induced velocity asymmetry, we fabricated two types of thin film structures, as detailed in Supporting Information, section S3. One structure has a relatively low DMI (LD), while the other has a relatively high DMI (HD). The LD structure consists of Si/SiO₂/Ta(25)/Pt(60)/CFB(7)/Pd(2)/Pt(16), and the HD structure comprises Si/SiO₂/Ta(25)/Pt(60)/CFB(6)/Pd(6)/Pt(12). Here, the numbers in parentheses indicate the thickness of each layer in angstroms, with CFB representing the Co₆₀Fe₂₀B₂₀ alloy. The thicknesses of the CFB layers were adjusted in both structures to achieve PMA with roughly similar anisotropy fields (Supporting Information, section S4). Here, an in-plane charge current is converted by the Pt layer into a perpendicular spin current and/or spin accumulation, which induces DW motion in the CFB layer via the spin–orbit torques (SOTs).^9,13−15,23 The role of the top Pd–Pt bilayer is to modulate the interfacial DMI of the CFB layer. In these types of structures, the DMI is given mainly by the bottom Pt/CFB interface, while the DMI of the top CFB/(Pd)/Pt interface will add destructively. By varying the thickness of the Pd layer, the iDMI induced by the top Pt layer can be gradually screened²⁴ enabling precise control over the total DMI. Brillouin light scattering (BLS) experiments (see Supporting Information, section S5) revealed an effective DMI constant of 0.48 mJ/m² for the LD structure and 1.1 mJ/m² for HD structure.

To control the charge current direction relative to the DW, the samples were patterned as 3 μm wide DW conduits on top of 26 μm wide Pt strips by using photolithography and Ar ion milling (see Supporting Information, section S3). We first analyze the current induced DW motion in a straight DW conduit, shown schematically in Figure 2(a). The DWs were nucleated and placed in their initial positions by an appropriate series of out-of-plane magnetic field pulses. Voltage pulses of varying amplitude and a width of 2.75 ns were applied on the Pt strip, and the resulting DW displacement was observed using wide-field magneto-optical Kerr effect (MOKE) microscopy. A voltage pulse of 25 V is equivalent to a current density of 1.2 ÷ 1.3 × 10¹² Am^–2, assuming that the current is flowing only through the current strip. This assumption overestimates the current through the Pt layer, providing only an upper bound. Although a more accurate estimation would require employing more sophisticated methods,²⁵ determining the exact current density is not critical for the objectives of our study. Figure 2(c)–(f) displays typical differential MOKE images obtained by subtracting the initial image from the one recorded after the application of the pulses. The dark contrast indicates the motion of an up/down (down/up) DW toward the left (right) side, while the light contrast shows the motion of an up/down (down/up) DW toward the right (left) side. The DWs move in the direction of the applied current, as expected for the left-handed DW chirality.¹³Figure 2(b) shows the DW velocity as a function of the pulse amplitude. Both types of samples show similar behavior. Above a certain threshold, the velocity increases rapidly, indicating the transition from the creep to the flow DW motion regime.²⁶ The onset of DW velocity saturation is also observed at a higher value for the HD sample compared to the LD one, as expected for the SOT-DMI current induced DW motion mechanism.^13,15,27

Figure 2.

(a) Schematic representation of a straight CFB DW conduit patterned on top of the Pt current strip. Red (blue) regions denote up (down) magnetized domains. An up/down DW is positioned in the middle of the conduit. (b) DW velocity as a function of pulse amplitude for HD and LD samples. Beyond a certain threshold, the velocity undergoes a rapid increase, characteristic of thermally activated DW motion, marking the transition to the flow DW motion regime. The onset of DW velocity saturation is also observed, occurring for higher values for the HD samples compared to the LD ones, consistent with the SOT-DMI current-induced DW motion mechanism. Points are average values of at least 10 measurements on four patterned devices, using two sets of each sample deposited in different runs. Error bars correspond to the standard deviation. (c)–(f) Differential MOKE images showing DW propagation after the application of current pulses. Dark contrast indicates up/down (down/up) DW motion to the left (right), while bright contrast indicates up/down (down/up) DW motion to the right (left). DWs move in the direction of the applied current, consistent with left-handed DW chirality. In (c), the DW conduit boundaries are highlighted by white dashed lines; the Pt current strip is shown by blue dashed lines, and the red shaded area marks the electrical contacts.

In the second step, the DW conduit was designed with a section forming an ±45° angle relative to the current direction (Figure 3(a)). This geometry allows the observation of DW motion within the same sample when the current is either parallel to or at a ±45° angle relative to the DW conduit. Figure 3(c) shows MOKE images indicating the displacement of a down/up DW under the action of positive and negative pulses in the LD sample. Initially, a down/up DW is nucleated in the lower left part of the structure, and a series of N = 280 positive pulses with an amplitude of 2.1 × 10¹²Am⁻² are applied (Figure 3(c), top). The DW moves under the action of the current through the entire conduit up to the right top side of the structure. Subsequently, the same number of pulses with amplitude 2.0 × 10¹² Am^–2 are applied in the negative direction. In this case, the DW motion is blocked in the 45° zone of the DW conduit. This occurs immediately after the DW passes the corner of the conduit, when the DW angle approaches 45° relative to the current strip (Figure 3(c), middle; see also Supporting Information Video SV1). To confirm that the DW is indeed blocked, a background MOKE image was recorded and extracted from the image taken after application of additional 10 × N pulses and shown in the bottom of part Figure 3(c). No discernible DW displacement is noticeable. For geometric reasons, a down/up DW in the 45° zone under positive pulses corresponds to the situation described in Figure 1(d-iv), while under negative pulses it corresponds to the one in Figure 1(b-iv). A similar situation occurs for an up/down DW (Figure 3(d)). The DW moves through the entire structure from the top right side to the bottom left side with the application of N = 280 negative pulses (Figure 3(d), top). When positive pulses are applied, the DW is blocked in the 45° zone (Figure 3(d), middle; see also Supporting Information Video SV2). Further application of positive pulses does not result in a discernible DW displacement (Figure 3(d), bottom). Similarly, an up/down DW in the 45° zone under positive pulses corresponds to the situation described in Figure 1(b-iv), while under negative pulses it corresponds to the one in Figure 1(d-iv).

Figure 3.

(a) Schematic representation of a CFB DW conduit patterned on the Pt current strip, with a section tailored to form a ±45° angle relative to the current direction. The red region indicates an up-magnetized domain, while blue region indicates a down-magnetized domain, with an up/down DW located in the middle of the 45° zone. (b) DW velocity in the 45° zone normalized to the velocity in the straight zone for the LD and HD samples. The points represent average values, with error bars extending to the most distant outliers. LD sample: (c) A down/up DW moving under the action of N = 280 positive (top part) and negative pulses (middle part). For positive pulses, the DW moves through the entire structure, and for negative pulses, the motion is blocked in the 45° zone, even after further application of 10 × N pulses (bottom part). (d) An up/down DW moving under the action of N = 280 negative (top part) and positive pulses (middle part), with a similar behavior as in (c). HD sample: (e) A down/up DW moving under the action of N = 175 positive (top part) and negative pulses (middle part). For negative pulses, the DW motion is not blocked in the 45° zone, but its velocity is diminished (bottom part). (f) An up/down DW moving under the action of N = 175 negative (top part) and positive pulses (middle part). For positive pulses, the DW motion is blocked in the 45° zone, and further application of N positive pulses does not set the DW in motion (bottom part).

For the HD sample, the DW behavior differs. A down/up domain moves through the entire structure from the lower left to the upper right with the application of N = 175 positive pulses (Figure 3(e), top). When the same number of negative pulses are applied, the DW reaches the 45° zone (Figure 3(e), middle; see also Supporting Information Video S3), but it is not blocked; its velocity is only diminished. Further application of N negative pulses results in a discernible DW displacement in the 45° zone (Figure 3(e), bottom; see also Supporting Information Video SV4). As in the case of the LD sample, an up/down domain nucleated in the top right side of the conduit will move through the entire structure when N = 175 negative pulses are applied (Figure 3(f), top). The DW motion is reversed for positive pulses and blocked in the 45° zone (Figure 3(f), middle part; see also Supporting Information Video SV5). Further application of negative pulses does not result in additional DW displacement (Figure 3(f), bottom; see also Supporting Information Video SV6).

Figure 3(b) shows the DW velocity in the 45° zone normalized to the velocity in the straight zone for the LD and HD samples. The points represent values averaged for up to four different patterned devices and for two series of each sample deposited in different runs, with error bars extended to include the furthest outliers. For the LD sample, the velocity in the 45° zone for a down/up DW moving to the right or an up/down DW moving to the left is slightly higher than the velocity in the straight zone. Conversely, a down/up DW moving to the left or an up/down DW moving to the right is always blocked in the 45° zone. For the HD samples, the velocity in the 45° zone for a down/up DW moving to the right or an up/down DW moving to the left is roughly the same as the velocity in the straight zone. However, the velocity of a down/up DW moving to the left or an up/down DW moving to the right is, on average, approximately 1 order of magnitude lower than the velocity in the straight zone. The fact that the normalized velocity in the 45° zone for a down/up DW moving to the right or an up/down DW moving to the left is slightly higher for the LD sample compared to the HD one aligns with our simple model and micromagnetic simulations, where a larger velocity asymmetry is expected for the lower DMI case for the same current density (see sections S1 and S2 of the Supporting Information). For a down/up DW moving to the left or an up/down DW moving to the right, the driving-force behind the DW motion decreases, and the DW velocity is largely determined by the pinning potential landscape and thermal activation.¹³⁻¹⁵ In the case of the LD lower DMI sample, the driving force decrease is large enough so no DW displacement can be detected even after application of 10 × N pulses. For the higher DMI HD sample, the DW moves fast between pinning sites, but experiences extended waiting periods at each site, resulting in an overall 1 order of magnitude decrease in velocity (see also Supporting Information, section S5). The observation that one type of DW can move unobstructed in one direction while being hindered and even blocked in the reverse direction demonstrates the realization of the DW diode functionality in the device.

Due to the specific geometry of our device, the out-of-plane Oersted field generated by the current passing through the Pt strip affects DW motion differently in the two straight sections. If it inhibits DW motion in the first straight section, it promotes motion in the second and vice versa, potentially influencing the overall functionality of the device. However, as demonstrated in Supporting Information, section S7, the effect of the out-of-plane Oersted field is negligible. Even if this were not the case, it would not significantly affect the functionality of the device as the out-of-plane Oersted field goes to zero at the center of the 45° zone and would not affect the DW velocity in that critical region.

An important question is whether the observed asymmetric behavior is intrinsic or an extrinsic material-dependent feature related to DW pinning. To address this, we deposited analogous samples using Co instead of CFB. Co has a higher depinning field, and its crystalline structure results in more defects compared to the amorphous CFB. Consequently, Co is expected to exhibit a higher density of pinning sites with stronger pinning potentials relative to CFB. As detailed in Supporting Information, section S7, the Co samples show a comportment analogous to that of the CFB samples. This supports the assumption that the observed asymmetric behavior is intrinsic rather than a material-dependent feature.

The DW velocity asymmetry between up/down and down/up DWs in the 45° zone can be exploited to construct DW selectors. To achieve this, we designed devices where the straight DW conduit splits into two branches oriented at −45° and 45° relative to the current direction. For the LD structure, a down/up domain is nucleated in the left part of the straight conduit, followed by the application of a series of N = 275 pulses with an amplitude of 2.1 × 10¹² Am^–2 (Figure 4(a)). As the DW reaches the bifurcation, it enters both branches. The DW motion is obstructed in the lower branch, where the DW conduit and current form a 45° angle. In contrast, the DW moves freely in the upper branch, where the DW conduit and current form a −45° angle, allowing it to reach the second straight zone. It is interesting to observe that the up/down DW behavior exhibits mirror symmetry (Figure 4(b)). In this case, the DW is blocked in the upper branch, where the DW conduit and current form a −45° angle and moves freely in the lower branch, where the DW conduit and current form a 45° angle. The HD structure shows a similar behavior with the application of N = 150 pulses (Figure 4(c) and (d)). The primary distinction is that the down/up (up/down) DW is not entirely blocked in the lower (upper) branch but progresses at a significantly reduced velocity (see also Supporting Information Video SV7). Consequently, our device demonstrates DW selective routing: down/up DWs traverse the upper branch freely, while up/down DWs traverse the lower branch freely.

Figure 4.

A Y-shaped DW selector with two branches oriented at −45° and 45° relative to the current direction. LD structure: (a) A down/up DW moving under the action of N = 275 positive pulses. The DW moves freely in the upper branch, while in the lower branch the DW motion is blocked. (b) An up/down DW showing mirror symmetry relative to (a), the DW is blocked in the upper branch and moves freely in the lower branch. HD structure: The down/up (c) or up/down (d) DWs show similar behavior with the application of N = 150 pulses, as in the in case of the LD structure. The only difference from the LD case is that the down/up (up/down) DW is not completely blocked in the lower (upper) branch, but it moves with reduced velocity.

Micromagnetic simulations (Supporting Information, section S2), supported by the experimental data, indicate that the velocity asymmetry depends on the DMI strength. Additionally, simulations reveal that adjusting the angles of the two output branches in the Y-shaped DW conduit can modulate this asymmetry. Specifically, the asymmetry increases as the angle between the branches increases, providing an additional means of controlling the DW velocity asymmetry. Consequently, the Y-shaped device can function as a DW splitter with asymmetric and tunable outputs, enabling precise control over DW trajectories in complex network architectures, analogous to in-plane magnetized DW networks.²⁸ This enhanced control over DW propagation could have significant implications for a diverse array of DW-based applications, including the development of unconventional computing schemes.²⁹

In summary, we have shown that adjusting the angle between the DW conduit and the current strip provides additional geometric control over the current-induced DW motion. Micromagnetic simulations highlighted the role of DL torque in creating an asymmetry in DW velocity when the current is applied at angles of 45° and −45° relative to the DW conduit. This velocity asymmetry was leveraged by designing the DW conduit with a 45° region. MOKE microscopy experiments demonstrated that both low-DMI (LD) and high-DMI (HD) structures exhibit DW diode-like behavior. For LD structures, a down/up (or up/down) DW moves freely through the entire structure under a series of positive (or negative) pulses, but reversing the current blocks the DW motion in the 45° zone. In HD samples, the behavior is similar except that the DW motion in the 45° zone is not completely blocked but rather has a significantly reduced velocity. Finally, we utilized the DW velocity asymmetry to construct DW selectors. In the Y-shaped DW conduit with one input and two, 45° and −45°, output branches, a DW injected into the junction will exit through one branch, while a DW of reverse polarity will exit through the other branch, thereby demonstrating selective routing of DWs.

Methods

Device Fabrication

The LD structure consisting of Si/SiO₂/Ta(25)/Pt(60)/CFB(7)/Pd(2)/Pt(16) and the HD structure comprising Si/SiO₂/Ta(25)/Pt(60)/CFB(6)/Pd(6)/Pt(12), where the numbers in parentheses indicate the thickness of each layer in angstroms and with CFB representing the Co₆₀Fe₂₀B₂₀ alloy, were deposited on thermally oxidized Si substrates using dc magnetron sputtering at an argon pressure of 1.5 mTorr. For current-induced DW motion experiments, the samples were patterned as 3 μm wide DW conduits on top of 26 μm wide Pt strips using UV photolithography and Ar-ion milling. The main steps of the device patterning are given in the Supporting Information.

Micromagnetic Simulations

For a better understanding of current-induced DW motion mechanisms, we use micromagnetic simulations performed with MuMax3 code. The system is discretized in 2 nm × 2 nm × 0.6 nm cells. The following magnetic parameters are used: saturation magnetization M_s = 1.37 × 10⁶ A/m, exchange coupling constant A = 10 × 10^–12 J/m, perpendicular uniaxial anisotropy constant K_u = 1.525 × 10⁶ J/m³, Gilbert damping parameter α_G = 0.6, spin-Hall angle θ_SHE = 0.07, and the ratio between the field-like and the damping like SOTs ξ = 0.2. The DMI was varied from |D| = 0.5 mJ/m² up to |D| = 1.1 mJ/m². Both ideal and disordered systems are considered.

MOKE Microscopy Measurements

The MOKE images were recorded by using a custom-built wide-field polar Kerr effect microscope. DW motion is driven by current pulses generated by a Picosecond Pulse Laboratories 2600C pulse generator, which supplies high-amplitude pulses up to 45 V with a fast rise time of less than 250 ps. The initial magnetic state of the device is prepared by the application of appropriate out-of-plane magnetic field pulses. Then, voltage pulses of varying amplitude and a width of 2.75 ns were applied within the Pt stripe, and the DW motion is observed using the MOKE microscope.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.nanolett.4c03339.

• Model of the DW magnetization dynamic tilting, micromagnetic simulations, details on the sample growth and microfabrication, magnetic properties and DMI characterization, tracking the DW motion in the 45° zone, and perpendicular Oersted field and ferromagnetic material effect (PDF)

• Movie SV1: MOKE movie demonstrating the diode functionality of the LD device. A down/up DW is displaced by the application of 280 positive and negative pulses (AVI)

• Movie SV2: MOKE movie demonstrating the diode functionality of the LD device. An up/down DW is displaced by the application of 280 negative and positive pulses (AVI)

• Movie SV3: MOKE movie demonstrating the diode functionality of the HD device. A down/up DW is displaced by the application of 175 positive and negative pulses (AVI)

• Movie SV4: MOKE movie showing the displacement of a down/up DW within the HD device by the application of 750 negative pulses. The DW is not blocked in the 45° zone, its velocity is reduced by approximately 1 order of magnitude (AVI)

• Movie SV5: MOKE movie demonstrating the diode functionality of the HD device. An up/down DW is displaced by the application of 175 negative and positive pulses (AVI)

• Movie SV6: MOKE movie showing the displacement of an up/down DW within the HD device by the application of 750 positive pulses. The DW is blocked in the 45° zone (AVI)

• Movie SV7: MOKE movie demonstrating the DW selector functionality of the HD device. Down/up and an up/down DWs are displaced by the application of 150 positive pulses. The down/up DW follows the upper branch, while the up/down DW follows the lower branch of the Y-shaped device (AVI)

Author Contributions

M.S.G. coordinated the research. E.M.S., T.P., and M.S.G. grew the films and patterned the devices. E.M.S., T.P., O.A.P., I.M.M., and M.S.G. conceived the experiments. E.M.S. performed the experiments. M.B. measured the DMI by BLS. M.S.G. supervised and wrote the manuscript with contributions from E.M.S. All the authors discussed the results and participated in preparing the manuscript. All authors have given approval to the final version of the manuscript.

This work was supported by a grant from the Romanian Ministry of Education and Research, CNCS-UEFISCDI, project number PN-III-P4-ID-PCE-2020-1853, within PNCDI III, and by CLOUDUT Project, cofounded by the European Regional Development Fund through the Competitiveness Operational Program 2014-2020, contract no. 235/2020. T.P. acknowledges funding from a grant from the Romanian Ministry of Education and Research, CNCS-UEFISCDI, project number PN-III-P1-1.1-TE-2021-1777, within PNCDI III. I.M.M. acknowledges funding for this work from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programs: ERC-StG Smart Design (638653) and ERC-PoC SOFT (963928).

The authors declare no competing financial interest.