MAGNETIC DIRECT ACCESS STORAGE-BASED COMPONENTS AND METHOD FOR MANUFACTURING THEM

Abstract

Magnetic direct access memory-based non-volatile Hall effect device comprising the following: a spin-orbit torque layer (10) arranged over a substrate (5); a magnetic layer (20) arranged above the spin-orbit torque layer (10); and a metal oxide layer (30) arranged above the magnetic layer (20), wherein sections of the spin-orbit torque layer (10) extend outwards from the magnetic layer (20) and the metal oxide layer (30) on opposite sides of a first direction and opposite sides of a second direction in a top view, the second direction being perpendicular to the first direction, wherein the spin-orbit torque layer (10) comprises heavy metals or materials doped with heavy metals or PtMn, wherein the heavy metals are selected from A and the materials doped with heavy metals are selected from M doped with B; where A contains: Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Cd, In, Hf, Re, Os, Ir, Au, Hg, Tl, Pb, Po, At and / or combinations thereof; where M contains at least one of the following: Al, Ti, Si, GaMn, GaAs; and where B has at least one of the following: Fe, Co, Ni, P, S, Ge, As, Se, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Cd, In, Sb, Te, I, Lu, Os, Ir, Bi, Po, At, La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb.

Description

GENERAL STATE OF THE ART

[0001] Magnetic random access memory (MRAM) offers performance comparable to volatile static random access memory (SRAM) and comparable density with lower power consumption than volatile dynamic random access memory (DRAM). Compared to flash memory with non-volatile memory (NVM), MRAM offers much faster access times and experiences less degradation over time, whereas flash memory has a limited number of rewrites. One type of MRAM is spin-transfer torque random access memory (STT-RAM). STT-RAM employs a magnetic tunnel junction (MTJ) that is at least partially written by a current driven through the MTJ. Another type of MRAM is spin-orbit torque RAM (SOT).

[0002] Prior art relating to the subject matter of the invention can be found, for example, in CN 1 08 321 292 A, CN 1 08 336 222 A, US 2018 / 0 123 021 A1, US 9 824 735 B1, WO 2017 / 034 563 A1, and DE 10 2013 109 012 A1. The invention is defined by the main claim and the subordinate claims. Further embodiments of the invention are recited in the dependent claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0003] This disclosure is best understood by reference to the following detailed description and accompanying drawings. It is emphasized that, in accordance with industry practice, various elements are not drawn to scale and are shown for illustrative purposes only. Indeed, the dimensions of various features may be arbitrarily exaggerated or reduced for clarity of explanation. Fig. 1 is a schematic illustration of an MRAM-based NVM Hall effect device according to an embodiment of the disclosure. Fig. 2 is a schematic cross-sectional view of an MRAM-based NVM Hall effect device according to an embodiment of the disclosure. Fig. 3 is a schematic cross-sectional view of an MRAM-based NVM Hall effect device according to an embodiment of the disclosure. Fig. 4 is a schematic cross-sectional view of an MRAM-based NVM Hall effect device according to an embodiment of the disclosure. Fig. 5 is a flow diagram illustrating a method of fabricating an MRAM-based NVM Hall effect device according to an embodiment of the disclosure. The Fig. 6A, Fig. 6B, Fig. 6C and Fig. 6D illustrate various stages of a manufacturing process for an MRAM-based NVM Hall effect device according to an embodiment of the present disclosure. The Fig. 7A, Fig. 7B, Fig. 7C and Fig. 7D illustrate one of the various stages of a manufacturing process for an MRAM-based NVM Hall effect device according to an embodiment of the present disclosure. The Fig. 8A, Fig. 8B and Fig. 8C illustrate one of the various stages of a manufacturing process for an MRAM-based NVM Hall effect device according to an embodiment of the present disclosure. The Fig. 9A, Fig. 9B and Fig. 9C illustrate one of the various stages of a manufacturing process for an MRAM-based NVM Hall effect device according to an embodiment of the present disclosure. The Fig. 10A and Fig. 10B illustrate an MRAM-based NVM Hall effect device according to an embodiment of the present disclosure. Fig. 11 illustrates the operation of an MRAM-based NVM Hall effect device according to an embodiment of the present disclosure. The Fig. 12A, Fig. 12B and Fig. 12C illustrate the operation of an MRAM-based NVM Hall effect device according to an embodiment of the present disclosure. Fig. 13 shows an MRAM-based non-volatile physically unclonable device according to an embodiment of the disclosure. Fig. Figure 14 shows the random flipping of an MRAM-based non-volatile physically unclonable device when a write current is applied. Fig. 15 is a flow diagram illustrating a method of manufacturing an MRAM-based device according to an embodiment of the disclosure. Fig. 16 is a flow diagram illustrating a method for writing to an MRAM-based device according to an embodiment of the disclosure. Fig. 17 is a flow diagram illustrating a method of manufacturing an MRAM-based device according to an embodiment of the disclosure. Fig. 18 illustrates the randomness of an MRAM-based non-volatile physically unclonable device according to an embodiment of the disclosure. The Fig. 19A and Fig. 19B illustrate writing to and reading from an MRAM-based device in an array according to embodiments of the disclosure. The Fig. 20A, Fig. 20B and Fig. 20C illustrate an MRAM-based device embedded in an integrated circuit. DETAILED DESCRIPTION

[0004] It should be understood that the following disclosure discloses many different embodiments or examples for implementing various features of the invention. Specific embodiments or examples of components and devices are described below to simplify the present disclosure. For example, the dimensions of features are not limited to the disclosed range or values, but may depend on process circumstances and / or desired properties of the component.The formation of a first feature over or on a second feature in the following description may also include embodiments in which the first and second features are formed in direct contact, and may also include embodiments in which additional features may be formed interposed between the first and second features such that the first and second features may not be in direct contact. Various features may be arbitrarily drawn at different scales for simplicity and clarity. In the accompanying drawings, some layers / features may be omitted for simplicity.

[0005] Furthermore, spatial reference terms such as "beneath," "under," "lower," "above," "upper," and the like may be used herein for ease of description to describe the relationship of one element or feature to one or more other elements or features as illustrated in the figures. The spatial reference terms are intended to encompass different orientations of the component during use or operation in addition to the orientation depicted in the figures. The component may be oriented differently (rotated 90° or at other orientations), and the spatial reference descriptors used herein will be interpreted accordingly. Additionally, the term "made of" can mean either "including" or "consisting of."Furthermore, in the subsequent manufacturing process, one or more additional operations may exist within / between the described operations, and the order of the operations may be changed. In the present disclosure, a phrase "one of A, B, and C" means "A, B, and / or C" (A, B, C, A and B, A and C, B and C, or A, B and C), and does not mean an element of A, an element of B, and an element of C, unless otherwise described.

[0006] STT-MRAM exhibits features such as non-volatile nature, compatibility with complementary silicon metal oxide semiconductor (Si-CMOS) technology, fast read and write speeds, high endurance and data retention, a relatively small bit cell size, and environmental resistance, making it the next disruptive technology for all CMOS integrated circuits (ICs) requiring memory. One high-value emerging application for STT-MRAM is a low-level cache for a central processing unit (CPU) or micro control unit (MCU), which offers attractive benefits of its non-volatility, including increased system speed and faster power-up. However, this application imposes a demanding demand on the memory's speed, particularly the write speed, which is much slower than the read speed.Additionally, the cache application for an MCU requires low power consumption, which is difficult for STT-MRAM because it requires significant current to change the magnetization state during the write operation. For current STT-MRAMs, write speed improvements via a foil stack and write system optimization, as well as write current reduction via stack optimization and CD reduction, may be stalled due to unavoidable performance trade-offs in endurance and data retention. New ideas, such as high-frequency assisted write, have been proposed, but these may not be feasible. There is a significant gap between the best reported STT-MRAM write speed and power and those required by cache applications, which could prove to be a showstopper.

[0007] In contrast, magnetic spin-orbital transfer (or torque) switching (SOT) is an emerging concept that has the potential to provide orders of magnitude improvement in write current and speed. SOT is considered a solution for high-speed, low-power cache applications.

[0008] Fig. 1 is a schematic view of an MRAM-based non-volatile memory (NVM) Hall effect device employing spin-orbit interaction in switching according to an embodiment of the present disclosure.

[0009] The MRAM-based NVM Hall-effect device includes a metallic bottom layer 10 as an active spin-orbit interaction (SOT) layer formed over a support layer 5. The MRAM-based NVM Hall-effect device further includes a first magnetic layer 20, which is a free magnetic layer or a data storage layer, disposed over the metallic bottom layer 10, a metal oxide layer 30 disposed over the first magnetic layer, and, in some embodiments, a second magnetic layer 40 as a reference layer disposed over the metal oxide layer 30. In some embodiments, the metal oxide layer is non-magnetic. In some embodiments, an upper conductive layer 50 is disposed as an electrode over the second magnetic layer 40.

[0010] The magnetic moment of the free layer 20 (the first magnetic layer) is switched using the spin-orbit interaction effect. In some embodiments, the magnetic moment of the first magnetic layer 20 is switched using only the spin-orbit interaction effect. In other embodiments, the magnetic moment of the first magnetic layer 20 is switched using a combination of effects. For example, the magnetic moment of the first magnetic layer 20 is switched using spin-transfer torque as a primary effect, which may be assisted by torque induced by the spin-orbit interaction. In other embodiments, the primary switching mechanism is torque induced by the spin-orbit interaction. In such embodiments, another effect, including but not limited to spin-transfer torque, may assist in switching.

[0011] The metallic bottom layer 10 is an active spin-orbit layer that exhibits a strong spin-orbit interaction and can be used to switch the magnetic moment of the first magnetic layer 20. The metallic bottom layer 10 is used to generate a spin-orbit magnetic field H. More specifically, a current driven in-plane through the metallic bottom layer 10 and the associated spin-orbit interaction can result in the spin-orbit magnetic field H. This spin-orbit magnetic field H is equivalent to the spin-orbit torque T upon magnetization, where T=-γ[M×H] in the first magnetic layer 20. The torque and the magnetic field are therefore interchangeably called a spin-orbit field and a spin-orbit torque. This reflects the fact that the spin-orbit interaction is the origin of the spin-orbit torque and the spin-orbit field.Spin-orbit torque occurs when a current is driven in-plane in the metallic bottom layer 10 and a spin-orbit interaction occurs. In contrast, spin-transfer torque is due to a current flowing perpendicular to a plane through the first magnetic layer 20, the non-magnetic spacer layer 30, and the second magnetic layer 40 (reference layer), which injects spin-polarized charge carriers into the first magnetic layer 20. The spin-orbit torque T can rapidly deflect the magnetic moment of the first magnetic layer 20 from its equilibrium state parallel to the easy axis. The spin-orbit torque T can flip the magnetization of the first magnetic layer 20 considerably faster than conventional STT torque with a similar maximum amplitude. In some embodiments, switching can be accomplished using spin-orbit torque.In other embodiments, a different mechanism, such as spin transfer, may be used to complete the switching. The generated spin-orbit field / spin-orbit torque may therefore be used in switching the magnetic moment of the first magnetic layer 20.

[0012] In some embodiments, the interaction of the metallic lower layer comprises the spin Hall effect. For the spin Hall effect, a current I+ is driven in the plane of the metallic lower layer 10 in a first direction (x-direction) (i.e., current in-plane, substantially in the xy-plane in Fig. 1). In other words, the current I+ is driven perpendicular to the stacked direction of the foils, including the metallic lower layer 10 and the first magnetic layer 20 (i.e., perpendicular to the normal of the surface, in Fig. 1 (the z-direction) vertical. Charge carriers that have spins with a particular orientation perpendicular to the current direction and to the surface normal (z-direction) accumulate on the surfaces of the metallic lower layer 10. A large portion of these spin-polarized carriers diffuses into the first magnetic layer 20 (the free layer). This diffusion results in the torque T on the magnetization of the first magnetic layer 20. Since torque on the magnetization, as explained above, is equivalent to the effective magnetic field on the magnetization, the spin accumulation equivalently results in the field H on the first magnetic layer 20. The spin-orbit field for the spin Hall effect is the cross product of the spin-orbit polarization and the magnetic moment of the first magnetic layer 20. The magnitude of the torque is therefore proportional to the in-plane current density I+ and spin polarization of the carriers.The spin Hall effect can occur when switching the stacked magnetic layer, which is shown in Fig. 1, can be used when the polarization induced by the spin Hall effect is parallel to the easy axis of the first magnetic layer 20. To obtain the spin-orbit torque T, the current pulse is driven in-plane through the metallic bottom layer 10. The resulting spin-orbit torque T acts against the damping torque, resulting in the switching of the magnetization of the first magnetic layer 20 in a manner analogous to conventional STT switching.

[0013] A combination of the anomalous Hall effect and the reverse spin Hall effect due to the passage of current I+ through the SOT layer generates a transverse voltage (Hall voltage) in the SOT layer. The transverse voltage is used as a signal / memory in some embodiments. As shown in Fig. As shown in Figure 1, in some embodiments, a voltmeter 90 is connected to portions of the SOT layer 10 that extend in a second direction (y-direction) substantially perpendicular (transverse) to the current flow direction (x-direction). The voltmeter measures the voltage on the edges of the SOT layer in the second direction. The measured voltage difference indicates the magnetic orientation of the magnetic layer 20.

[0014] As stated above, the metallic bottom layer 10 is an active spin-orbit layer that provides a strong spin-orbit interaction with the first magnetic layer 20 (the free layer). In some embodiments, the SOT layer 10 comprises heavy metals or materials doped with heavy metals. Such materials can be selected from A and M doped by B. A comprises Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Cd, In, Sb, Te, Hf, Ta (including highly resistive amorphous β-Ta), W (including α-W and β-W), Re, Os, Ir, Pt, Au, Hg, Tl, Pb, Bi, Po, At, and / or combinations thereof; M comprises at least one of Al, Ti, V, Cr, Mn, Cu, Zn, Ag, Hf, Ta, W, Re, Pt, Au, Hg, Pb, Si, Ga, GaMn or GaAs, and B comprises at least one of V, Cr, Mn, Fe, Co, Ni, P, S, Zn, Ga, Ge, As, Se, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Ag, Cd, In, Sb, Te, I, Lu, Hf, Ta, W, Re, Os, Ir, Pt, Au, Hg, Tl, Pb, Bi, Po, At, La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb.In some embodiments, the SOT layer comprises Ir-doped Cu and / or Bi-doped Cu. The doping is in the range of about 0.1 to about 10 atomic percent. In other embodiments, the SOT layer is one or more of platinum, tungsten, tantalum, and PtMn. A thickness of the metallic lower layer 10 is in a range of about 2 nm to about 20 nm in some embodiments, and in a range of about 5 nm to about 15 nm in other embodiments. In some embodiments, an antiferromagnetic layer, made of IrMn, for example, is disposed between the metallic lower layer 10 and the support layer 5.

[0015] In this disclosure, an “element layer” or a “compound layer” generally means that the content of the element or compound is more than 99%.

[0016] The first magnetic layer 20 as a data storage layer is a free layer having a switchable magnetic moment. In some embodiments, the magnetic layer 20 is ferromagnetic and comprises one or more of Fe, Ni, and Co. In some embodiments, the magnetic layer 20 comprises CoxFeyB. The values ​​of x and y can vary to obtain different magnetic / crystalline properties. In some embodiments, the magnetic layer 20 has a coercivity greater than about 500 gauss, where 1 gauss corresponds to 10-4 T. In some embodiments, the magnetic layer 20 has a coercivity of greater than about 500 gauss to about 750 gauss. In some embodiments, the magnetic layer has a coercivity of greater than about 500 gauss to about 600 gauss.

[0017] In some embodiments, the first magnetic layer 20 includes a lower magnetic layer 22, a middle magnetic layer 25, and an upper magnetic layer 24, as shown in Fig. 2. In some embodiments, the bottom magnetic layer 22 is a cobalt-iron-boron (CoFeB) layer, a cobalt / palladium (CoPd) layer, and / or a cobalt-iron (CoFe) layer, which in some embodiments has a thickness in a range of about 0.6 nm to about 1.2. In certain embodiments, the bottom magnetic layer 22 is FexCoyB1-xy, where 0.50 ≤ x ≤ 0.70 and 0.10 ≤ y ≤ 0.30. In other embodiments, 0.55 ≤ x ≤ 0.65 and 0.15 ≤ y ≤ 0.25.

[0018] In some embodiments, the upper magnetic layer 24 is a cobalt-iron-boron (CoFeB) layer, a cobalt / palladium (CoPd) layer, and / or a cobalt-iron (CoFe) layer, which in some embodiments has a thickness in a range of about 1.0 nm to about 1.2 nm, or a NiFe layer, which has a thickness in a range of about 0.4 nm to about 3.0 nm. In certain embodiments, the upper magnetic layer 24 is Fe x Co y B 1-x-y , where 0.50 ≤ x ≤ 0.70 and 0.10 ≤ y ≤ 0.30. In other embodiments, 0.55 ≤ x ≤ 0.65 and 0.15 ≤ y ≤ 0.25. In some embodiments, the upper magnetic layer 24 is made of the same material as the lower magnetic layer 22. In other embodiments, the upper magnetic layer 22 is made of a different material than the lower magnetic layer 22.

[0019] The middle non-magnetic layer 25 is, in some embodiments, a coupling layer and is made of one or more of W, Mo, Pt, and Ru, and alloys thereof. The thickness of the middle non-magnetic layer 25 is, in some embodiments, in a range from about 0.2 nm to about 0.5 nm. The coupling material of the middle non-magnetic layer 25 couples the lower magnetic layer 22 and the upper magnetic layer 24 of the first magnetic layer 20. Such a coupling would break the symmetry, and therefore field-free switching is possible. As shown in Fig. 2, the magnetic field direction of the lower magnetic layer 22 is substantially horizontal (perpendicular to the film stacking direction), while in some embodiments, the magnetic field direction of the upper magnetic layer 24 is substantially vertical (parallel to the film stacking direction). In some embodiments, the magnetic field direction of, for example, the lower magnetic layer 22 may tilt at a small angle (e.g., 1 to 30 degrees), which may cause deterioration of a tunneling magnetoresistance (TMR) ratio. In some embodiments, the TMR ratio may be controlled by controlling the strength of at least one of the lower magnetic layer 22 and the upper magnetic layer 24.

[0020] In some embodiments, as in Fig. 3, interfacial layers 26 and 28 are disposed between the lower magnetic layer 22 and the middle non-magnetic layer 25, and between the middle non-magnetic layer 25 and the upper magnetic layer 24, respectively. In some embodiments, the interfacial layers 26 and 28 are made of FeB. The thickness of the interfacial layers 26 and 28 ranges from about 0.5 nm to about 2.0 nm in some embodiments. When the middle non-magnetic layer 25 is disposed between the lower magnetic layer 22 and the upper magnetic layer 24 (in other words, when the middle non-magnetic layer 25 is inserted into the first magnetic layer 20), a dead layer may be observed at the interface between the middle non-magnetic layer 25 and the lower magnetic layer 22 and / or the upper magnetic layer 24, which may weaken perpendicular magnetic anisotropy (PMA).By inserting the interface layers 26 and 28 it is possible to eliminate the dead layers and maintain or improve the PMA.

[0021] The non-magnetic metal oxide layer 30 is made of a dielectric material. In some embodiments, the non-magnetic metal oxide layer 30 comprises a crystalline or amorphous magnesium oxide (MgO) layer. In other embodiments, the non-magnetic metal oxide layer 30 is made of aluminum oxide. In some embodiments, the metal oxide layer 30 has a thickness in a range of about 0.3 nm to about 2 nm, and in other embodiments, the thickness of the non-magnetic layer 30 is in a range of about 0.5 nm to about 1.0 nm. The metal oxide layer 30 is a material required for the perpendicular magnetic anisotropy of the magnetic layer 20 in some embodiments.

[0022] In some embodiments, the device includes a second magnetic layer 40. The second magnetic layer 40 is a reference layer whose magnetic moment does not change. In some embodiments, the second magnetic layer 40 is made of the same material as the first magnetic layer 20, as set forth above. In some embodiments, the second magnetic layer 40 includes multiple layers of magnetic materials. In some embodiments, the second magnetic layer 40 includes a multilayer structure of cobalt (Co) and platinum (Pt). In some embodiments, a thickness of the second magnetic layer 40 is in a range of about 0.2 nm to about 1.0 nm, and in other embodiments, in a range of about 0.3 nm to about 0.5 nm.

[0023] In some embodiments, the second magnetic layer 40 is a multilayer comprising a synthetic antiferromagnetic layer having ferromagnetic layers separated by a non-magnetic layer, such as Ru. In some embodiments, a pinning layer, such as an antiferromagnetic layer that fixes the magnetic moment of the second magnetic layer 40 in place, is disposed over the second magnetic layer 40 with a Ru layer sandwiched between them. The first and second magnetic layers are crystalline in some embodiments.

[0024] In some embodiments, the device includes an upper conductive layer 50 as an electrode. The upper conductive layer includes one or more layers of Ta, Ru, Au, Cr, Pt, Cu, Ni, W, and Al.

[0025] The support layer 5 is made of a dielectric material, such as silicon oxide, silicon oxynitride, silicon nitride, aluminum oxide, magnesium oxide, or another suitable material. In some embodiments, the support layer 5 is a shallow trench isolation layer, an interlayer dielectric (ILD) layer, or an inter-metal dielectric (IMD) layer in a semiconductor device.

[0026] Continue as in Fig. As shown in Figure 2, in some embodiments, a metallic interlayer 60 is disposed between the non-magnetic metal oxide layer 30 and the second magnetic layer 40. In some embodiments, the metallic interlayer 60 is made of a non-magnetic material. In certain embodiments, the metallic interlayer 60 is made of Mg. A thickness of the metallic interlayer 60 is in a range of about 0.1 nm to about 0.6 nm in some embodiments, and in a range of about 0.2 nm to about 0.5 nm in other embodiments. In other embodiments, no metallic interlayer is used.

[0027] In some embodiments, an antiferromagnetic layer 70 is formed over the second magnetic layer, and a third magnetic layer 80 is formed over the antiferromagnetic layer 70, as shown in Fig. 2. The antiferromagnetic layer 70 helps fix the magnetic moment of the second magnetic layer 40. In some embodiments, the antiferromagnetic layer 70 comprises ruthenium (Ru) or another suitable antiferromagnetic material. In some embodiments, the thickness of the antiferromagnetic layer 70 is in a range of about 0.2 nm to about 0.8 nm.

[0028] The third magnetic layer 80 includes one or more layers of magnetic materials. In some embodiments, the third magnetic layer 80 includes one or more of cobalt, iron, nickel, and platinum. In some embodiments, the material of the third magnetic layer 80 is the same as or different from the material of the second magnetic layer 40. In certain embodiments, the third magnetic layer 80 is a CoPt layer. A thickness of the third magnetic layer is in a range of about 0.5 nm to about 1.5 nm in some embodiments, and in a range of about 0.7 nm to about 1.2 nm in other embodiments.

[0029] Each of the layers in the Fig. 1 to 3 may be formed by any suitable film formation process including physical vapor deposition (PVD) including sputtering; molecular beam epitaxy (MBE); pulsed laser deposition (PLD); atomic layer deposition (ALD); electron beam (e-beam) epitaxy; chemical vapor deposition (CVD) or derivative CVD processes further including low pressure CVD (LPCVD), ultra high vacuum CVD (UHVCVD), reduced pressure CVD (RPCVD); electroplating, or any combinations thereof.

[0030] In some embodiments, a foil stack is formed by the above-mentioned foil forming processes, and after forming the foil stacks, a patterning process comprising one or more lithography and etching processes is performed on the foil stack to form an SOT cell as in Fig. 1 shown.

[0031] In some embodiments, a metallic bottom layer 10 is formed over a support layer 5. The metallic bottom layer 10 can be formed using PVD, CVD, ALD, or any other suitable film formation method. The metallic bottom layer 20 can be formed using PVD, CVD, ALD, or any other suitable film formation method. As set forth above, the first magnetic layer 20 includes the bottom magnetic layer 22, the middle non-magnetic layer 25, and the top magnetic layer 24. In some embodiments, the first magnetic layer 20 further includes the interfacial layers 26 and 28. These layers are formed sequentially over the metallic bottom layer 10. Further, the remaining layers are formed sequentially over the first magnetic layer 20.

[0032] Fig. Figure 4 is a schematic cross-sectional view of an SOT MRAM cell according to an embodiment of the present disclosure. In this embodiment, the order of the stacked foils is reversed.

[0033] Embodiments of the present disclosure combine a Hall sensor with a magnetic random access memory (MRAM) to provide a non-volatile memory (NVM) device. Unlike a Hall sensor, which requires an external magnetic field, embodiments of the present disclosure do not require an external magnetic field. The present disclosure provides a non-volatile Hall effect device that is less complicated than conventional Hall effect devices.

[0034] A method for manufacturing an MRAM-based NVM Hall effect device is described with reference to Fig. 5 to 9C. Fig. 5 is a flowchart illustrating a method of fabricating an MRAM-based NVM Hall-effect device according to an embodiment of the disclosure. At operation S210, a spin-orbit torque layer 10 is formed over a substrate 5. Then, a magnetic layer 20 is formed over the spin-orbit torque layer 10 at operation S220. A metal oxide layer is subsequently formed over the magnetic layer at operation S230. At operation S240, the spin-orbit torque layer 10, the magnetic layer 20, and the metal oxide layer 30 are patterned such that, in plan view, portions of the spin-orbit torque layer 10 extend outward from the magnetic layer 20 and the metal oxide layer 30 on opposite sides of a first direction and opposite sides of a second direction, where the second direction is perpendicular to the first direction.

[0035] The method of manufacturing an MRAM-based device according to some embodiments of the disclosure is further described with reference to the Fig. 6A to 9C. As described in Fig. As shown in Figure 6A, in some embodiments, an insulating layer 15 is formed over a semiconductor substrate 5. In some embodiments, the semiconductor substrate 5 is a silicon substrate. In some embodiments, the substrate 5 is a silicon wafer, and the insulating layer 15 is a silicon oxide layer. In some embodiments, the insulating layer 15 is formed by a CVD or PVD process; in other embodiments, the insulating layer is formed by thermally oxidizing the substrate 5.

[0036] In some embodiments, lower electrodes 45 are formed along a portion of the substrate 5 and the insulating layer 15, as shown in Fig. 6B. The insulating layer 15 is formed by photolithography and etching processes, in some embodiments followed by the deposition of an electrically conductive material. In some embodiments, excess electrically conductive material is removed by a polishing or etch-back process. In some embodiments, a chemical mechanical polishing (CMP) process is performed to planarize the top surface of the lower electrodes 45 and the insulating layer 15 or the substrate 5. In some embodiments, the electrically conductive material is selected from Ta, Ru, Au, Cr, Pt, W, Cu, Ni, Al, and alloys thereof.

[0037] In some embodiments, an SOT layer 10 is then deposited over the insulating layer 15 and the lower electrodes 45, as shown in Fig. 6C. In some embodiments, the SOT layer 10 is formed from one or more of platinum, tungsten, tantalum, and PtMn.

[0038] A magnetic layer 20 is formed over the SOT layer 10, and a metal oxide layer 30 is formed over the magnetic layer 20, as shown in Fig. 6D. In some embodiments, the magnetic layer Co x Fe y B, and the metal oxide layer 30 is MgO to form a layer stack. In some embodiments, the layer stack includes one or more additional magnetic layers 40 formed over the metal oxide layer 30 to modify the magnetic properties (Hc) of the magnetic layer 20.

[0039] The layer stack is then, as shown in the Fig. 7A, Fig. 7B and Fig. 7C shown, structured. Fig. 7A is a plan view, Fig. Figure 7B is a cross-sectional view along line AA' in Fig. 7A taken, and Fig. Figure 7C is a cross-sectional view taken along line BB' in Fig. 7A. As shown, the second magnetic layer 40, the metal oxide layer 30, the first magnetic layer 20, and the SOT layer 5 are patterned to form a cross shape, although the present disclosure is not limited to this shape. The layer stack is patterned using photolithography and etching processes.

[0040] Then, the second magnetic layer 40, the metal oxide layer 30 and the first magnetic layer 20 are further patterned to adjust their widths relative to the SOT layer 10 as shown in the Fig. 8A, Fig. 8B and Fig. 8C shown. Fig. 8A is a plan view, Fig. 8B is a cross-sectional view taken along line AA', and Fig. Figure 8C is a cross-sectional view taken along line BB' in Fig. 8A. The second magnetic layer 40, the metal oxide layer 30, and the first magnetic layer 40 are further patterned using photolithography and etching processes. The optional second magnetic layer 40, the metal oxide layer 30, and the first magnetic layer 20 are patterned such that these patterned layers overlie a central portion of the cross-shaped SOT layer 10. In other words, opposite sides of the SOT layer 10 extend outward from the edges of the first magnetic layer 20, the metal oxide layer 30, and the optional second magnetic layer 40 in substantially perpendicular directions. The bottom electrodes 75, in some embodiments, are electrically connected to the four ends of the cross-shaped SOT layer 10.

[0041] A top contact 50 comprising an electrically conductive material is subsequently formed on the MRAM-based device, as shown in FIGS. Fig. 9A, Fig. 9B and Fig. 9C shown. Fig. 9A is a plan view, Fig. 9B is a cross-sectional view taken along line AA' in Fig. 9A taken, and Fig. Figure 9C is a cross-sectional view taken along line BB'. The top contact may be formed using photolithography and material deposition techniques.

[0042] Fig. 10A is a cross-sectional view of an MRAM-based NVM Hall effect device according to embodiments of the disclosure, and Fig. Figure 10B is a schematic isometric view of the device without the substrate, showing the operation of the device before a write current is applied.

[0043] Fig. 11 is a schematic isometric view of the MRAM-based NVM Hall effect device of the Fig. 10A and Fig. 10B, which shows the operation of the device when the magnetic alignment is initialized. The initial magnetic alignment is indicated by the downward-pointing arrow 125. The initial magnetic alignment is set by applying either a magnetic field 120 ranging from about 100 Oe (Oersted) to about 2000 Oe, or a high current 115. In some embodiments, the magnetic field ranges from about 200 Oe to about 1000 Oe. In some embodiments, the applied current 115 is applied to the SOT layer 10 at a current density greater than 1×10 11A / m2 is applied along the x-direction. In some embodiments, current 115 is a DC pulse applied for a duration of at least 1 ns. In some embodiments, DC pulse 115 is applied for a duration of 1 ns to 1 s. In some embodiments, the DC pulse is applied for a duration of 1 ns to 1 ms. The transverse voltage is measured across the y-direction by a voltmeter 90 connected to ends of SOT layer 10 to determine the magnetic orientation of magnetic layer 20.

[0044] The Fig. 12A to 12C are schematic isometric views of the MRAM-based NVM Hall effect device. Fig. Figure 12A is a schematic view illustrating the application of a write current I+ to the MRAM-based device of the Fig. 11 shows. The write current I+ is applied to the SOT layer 10 along the x-direction, and the transverse voltage is measured by the voltmeter 90 across the y-direction from portions of the SOT layer 10 that extend beyond the magnetic layer 20 along the y-direction. If, as in Fig. 12B, the write current I+ is smaller than the threshold current of the particular device, the magnetic orientation 125 of the device does not change from its initial state, while in Fig. 12C, when the write current is greater than the threshold, the magnetic orientation of the device is reversed. Whether the magnetic orientation 125 has changed is checked by observing the voltage change using voltmeter 90. The magnetic orientation depends on the current. In some embodiments of SOT-MRAMs, the magnetic orientation does not return from a second to a first magnetic orientation upon the application of additional current pulses greater than the threshold current in the same current direction.

[0045] In some embodiments, the magnetic alignment of the magnetic layer is achieved by applying a current density of at least about 1×10 11 A / m2 to the SOT layer. In some embodiments, the current density applied to the SOT layer ranges from about 1×10 11 A / m2 to about 1×10 12A / m2. The invention is not limited to these current densities, since the value of the current density required to change the magnetic orientation depends on the properties of the materials of the device components.

[0046] In some embodiments, a plurality of MRAM-based devices (or cells) are configured in an array. Depending on the processing temperature of the plurality of MRAM-based devices and the thickness of the individual layers of the MRAM-based device, including the tunnel junction layer, the magnetic properties of the MRAM-based devices vary among the individual devices (cells) in the array. For example, a particular write current near a threshold voltage applied to a line or array of devices may cause some devices to reverse their magnetic orientation while other devices in the array do not switch. Whether a particular device changes its magnetic orientation is random and based on local differences in the layers of each device. The coercivity is randomly distributed within a certain range.Some components can therefore change their magnetic orientation at a particular write current, while other components do not change their magnetic orientation.

[0047] This naturally occurring randomness of coercivity allows a plurality of MRAM-based devices according to the present disclosure to be used in physically unclonable function (PUF) applications. In one embodiment, the devices in an array 100 are initialized (set to 0) by applying a current I+ 130 above the threshold current of each of the cells to the SOT layer 10. In some embodiments, an external magnetic field is used to initialize the plurality of devices in the array. As shown in Fig. 13, the magnetic alignment 125 of each individual component is set in the same direction. Then, Fig. 14, a write current I+ 135 close to the threshold current is applied to the array. Since the write current I+ 135 applied to the devices is close to the threshold, some devices flip from 0 to 1, while others remain in the 0 state, as shown in Fig. 14. Which particular device in an array will flip is random and unpredictable. The devices are therefore physically unclonable. The randomness, in some embodiments, is a result of the manufacturing process. In some embodiments, the physically unclonable feature is used to create a random structure in the array 100. In some embodiments, this random structure is used as electronic identification. In some embodiments, these random structures can be used as a security feature, such as a magnetic stripe on a credit card.

[0048] In some embodiments, the write current I+ 135 is varied to vary the number of devices that change magnetic orientation. Since the write current I+ 135 is close to the threshold current, reducing the write current I+ 135 applied to the spin-orbit torque layer 10 reduces the number of devices in which the magnetic orientation of the magnetic layer 20 is changed. Conversely, increasing the write current increases the number of devices in which the magnetic orientation of the magnetic layer 20 is changed.

[0049] In some embodiments, the array is a linear array of cells 100, as in Fig. 13 and Fig. 14. The cells are interconnected by the spin-orbit torque layer 10, and the linear array extends along the direction of flow of the write current I+. The linear array of cells 100 is formed, in some embodiments, by material deposition, photolithography, and etching processes as described herein. In other embodiments, a plurality of linear arrays of MRAM-supported cells are configured to form a two-dimensional array having a plurality of rows or columns of MRAM-supported cells, as viewed in plan view.

[0050] Fig. 15 is a flowchart illustrating a method of fabricating an MRAM-based device according to an embodiment of the disclosure. At operation S310, a spin-orbit torque layer 10 is formed over a substrate 5. A first magnetic layer 20 is formed over the spin-orbit torque layer 10 at operation S320. Then, a metal oxide layer 30 is formed over the magnetic layer 20 at operation S330. In some embodiments, a second magnetic layer 40 is subsequently formed over the metal oxide layer 30 at operation S340.In operation S350, the spin-orbit torque layer 10, the first magnetic layer 20, the metal oxide layer 30, and the optional second magnetic layer 40 are patterned such that portions of the spin-orbit torque layer 10 extend outward from the first magnetic layer 20, the metal oxide layer 30, and the second magnetic layer 40 on opposite sides of a first direction and opposite sides of a second direction in plan view, where the second direction is perpendicular to the first direction. In some embodiments, an intermediate layer 60 is formed in operation S360 after the metal oxide layer 30 is formed and before the second magnetic layer 40 is formed. In some embodiments, a conductive layer 50 is formed over the second magnetic layer 40 in operation S370.

[0051] Fig. 16 is a flow diagram illustrating a method for writing an MRAM-based device according to an embodiment of the disclosure. At operation S410, a magnetic orientation of a first magnetic layer of the MRAM-based device is determined. The device includes a spin-orbit torque (SOT) layer 10 disposed over a substrate 5. A first magnetic layer 20 is disposed over the spin-orbit torque layer 10, and a metal oxide layer 30 is disposed over the first magnetic layer 20. In some embodiments, the magnetic orientation is determined by measuring the transverse strain across portions of the SOT layer extending from the magnetic layer of the device in directions substantially perpendicular to the direction of current flow in the SOT layer. At operation S420, a current density of at least 10 11A / m2 is applied to the spin-orbit torque layer 10. In some embodiments, operation S430 determines whether the magnetic orientation of the first magnetic layer 20 changes after applying a current density of at least 10 11 A / m2 to a spin-orbit torque / layer has changed. In some embodiments, it is determined whether the magnetic orientation of the first magnetic layer 20 has changed by applying a read current to the SOT layer of the MRAM-based device at operation S440. In some embodiments, determining the magnetic orientation of the first magnetic layer comprises applying a read current to the SOT layer of the MRAM-based non-volatile Hall effect device at operation S450.

[0052] Fig. 17 is a flow diagram illustrating a method of fabricating an MRAM-based device according to an embodiment of the disclosure. In operation S510, a spin-orbit torque layer 10 is formed over a substrate 5. The spin-orbit torque layer 10 is patterned to form a cross-shaped spin-orbit torque layer over the substrate 5 in operation S520. In operation S530, a first magnetic layer 20 is formed over a central portion of the cross-shaped spin-orbit torque layer 10, and a metal oxide layer 30 is formed over the first magnetic layer 20 in operation S540. Portions of the spin-orbit torque layer 10 extend outward from the first magnetic layer 20 and the metal oxide layer 30 on opposite sides of a first direction and opposite sides of a second direction in a plan view, where the second direction is perpendicular to the first direction.In some embodiments, a second magnetic layer 40 is formed over the first magnetic layer 20 at operation S550. In some embodiments, a mask having an opening is formed over a central portion of the spin-orbit torque layer, and the second magnetic layer is formed in the mask opening. In some embodiments, an intermediate layer 60 is formed between the formation processes of the first magnetic layer 20 and the second magnetic layer 40 at operation S560. In some embodiments, a mask having an opening is formed over a central portion of the spin-orbit torque layer 10, and the intermediate layer 60 is formed in the mask opening.In some embodiments, a mask layer is formed over the spin-orbit torque layer before forming the first magnetic layer at operation S570, and an opening is formed in the mask layer at operation S580 by exposing the central portion of the spin-orbit torque layer. In some embodiments, the mask layer comprises a photoresist layer. In some embodiments, the metal oxide layer 30 comprises a material that maintains a perpendicular magnetic anisotropy of the first magnetic layer 20. In some embodiments, an electrically conductive layer 50 is formed over the first magnetic layer 20 at operation S590.

[0053] Fig. Figure 18 shows the randomness in a 10 x 8 array of 600 MRAM-based devices upon application of a write current close to the threshold current. Devices that change magnetic orientation are called "1," and devices that do not change orientation are called "0." Due to slight variations in the individual SOT MRAM cells, such as variations in layer thickness, a random pattern of cells that change magnetic orientation and cells that do not change orientation is generated. The array can therefore be used in physically unclonable function (PUF) applications, such as electronic identification. In some embodiments, these random patterns can be used as a security feature, such as a magnetic stripe on a credit card.

[0054] Fig. 19A illustrates a portion of the wiring of a circuit 700 including an array of MRAM-based devices according to an embodiment of the disclosure. The devices are connected to word lines WL, source lines SL, and bit lines BL. As shown, a write current is applied to the device from source line SL1 to word line WL1. The transverse (Hall) voltage signal is measured on bit line BL1. In some embodiments, the portion of spin-orbit torque layer 10 opposite the V output is voltage-free. The charge accumulates on the voltage-free side of the SOT layer. In other embodiments, the SOT layer has no floating. In some embodiments, another bit line BL1, BL2 is connected to the opposite side of the spin-orbit torque layer from the V output, as shown in Fig. 19B. In some embodiments, connecting opposite sides of the spin-orbit torque layer to two bitlines (e.g., BL1+ and BL1-) provides improved signals.

[0055] In some embodiments, the Hall voltage signal ranges from about 3% to about 50% of the applied current. The value of the Hall voltage signal depends on the SOT material. When a specific word line WL is activated, two transistors of a cell are turned on. An input (or write) current then flows through the spin-orbit torque layer of the MRAM-based device, generating a Hall voltage Vreadout. The Hall voltage Vreadout appears on the bit line as the cell's data. For example, activating WL1 and SL1 causes a voltage output Vreadout at the 11 (WL1 and BL1) position. Activating WL1 and BL2 causes a voltage output Vreadout at the 12 (WL1 and BL2) position, and so on.

[0056] The Fig. 20A, Fig. 20B and Fig. 20C illustrate an embodiment of an MRAM-based device embedded in an integrated circuit 800. As in Fig. 20A, the device 810 is located at the M2 wiring level above the source line SL, bit line BL, and word line WL of the integrated circuit 800. The device is connected by electrically conductive contacts 105 to the source line SL and the source / drain regions 75 of the transistors 115, 115' below the M1 wiring level. The transistors 115, 115' include a gate electrode 85 and a gate dielectric layer 95 above a channel region 65 between the source / drain regions 75 formed in a semiconductor substrate 55. In some embodiments, the gate electrodes 85 of adjacent transistors 115, 115' are connected by the word line WL. Fig. Figure 20B is an isometric (three-dimensional) view of the MRAM-based device of the Fig. 20A, and Fig. Figure 20C is a top view of the MRAM-based device of the Fig. 20A.

[0057] MRAM-based devices according to the present disclosure can provide an order of magnitude improvement in write current and speed. Devices according to the present disclosure are advantageous for high-speed, low-power cache applications. In some embodiments, MRAM-based non-volatile memory Hall sensor switches are formed. The MRAM-based non-volatile memory Hall sensor switches do not require an external magnetic field. In some embodiments, the magnetic non-volatile memory of the devices according to the disclosure retains its magnetic alignment for about 10 years or more. The naturally occurring randomness of the coercivity of devices according to the present disclosure is advantageous in physically unclonable function (PUF) applications, including electronic identification applications.

[0058] One embodiment of the disclosure is a magnetic random access memory-based non-volatile Hall effect device comprising a spin-orbit torque layer disposed over a substrate and a magnetic layer disposed over the spin-orbit torque layer. A metal oxide layer is disposed over the magnetic layer. Portions of the spin-orbit torque layer extend from the magnetic layer and the metal oxide layer on opposite sides of a first direction and opposite sides of a second direction in a plan view, the second direction being perpendicular to the first direction. In one embodiment, the metal oxide layer comprises a material that maintains a perpendicular magnetic anisotropy of the magnetic layer. In one embodiment, the magnetic layer has a coercivity greater than about 500 Gauss. In one embodiment, the metal oxide layer comprises MgO.In one embodiment, the magnetic layer comprises a CoFeB material. In one embodiment, the spin-orbit torque layer comprises one or more of platinum, tungsten, tantalum, and PtMn. In one embodiment, the substrate comprises a semiconductor substrate. In one embodiment, the substrate comprises silicon. In one embodiment, the substrate comprises a semiconductor substrate with an insulating layer interposed between the semiconductor substrate and the spin-orbit torque layer. In one embodiment, the spin-orbit torque layer is cross-shaped in plan view, and the magnetic layer and the metal oxide layers are disposed over a central portion of the cross-shaped spin-orbit torque layer. In one embodiment, magnetic alignment of the magnetic layer is achieved by applying a current density of 10 11A / m2 to the spin-orbit torque layer. In one embodiment, the device maintains its magnetic alignment for 10 years without the application of an external current or magnetic field. In one embodiment, the device includes a second magnetic layer disposed over the metal oxide layer.

[0059] One embodiment of the disclosure is a method of fabricating a magnetic random access memory-based non-volatile Hall effect device, comprising forming a spin-orbit torque layer over a substrate and forming a magnetic layer over the spin-orbit torque layer. A metal oxide layer is formed over the magnetic layer. The spin-orbit torque layer, magnetic layer, and metal oxide layer are patterned such that, in plan view, portions of the spin-orbit torque layer extend outwardly from the magnetic layer and the metal oxide layer on opposite sides of a first direction and opposite sides of a second direction, wherein the second direction is perpendicular to the first direction. In one embodiment, the metal oxide layer comprises a material that maintains a perpendicular magnetic anisotropy of the magnetic layer. In one embodiment, the metal oxide layer comprises MgO.In one embodiment, the magnetic layer has a coercivity greater than about 500 gauss. In one embodiment, the magnetic layer comprises a CoFeB material. In one embodiment, the spin-orbit torque layer comprises one or more of platinum, tungsten, tantalum, and PtMn. In one embodiment, the substrate comprises a semiconductor substrate. In one embodiment, the substrate comprises silicon. In one embodiment, the substrate comprises a semiconductor substrate with an insulating layer interposed between the semiconductor substrate and the spin-orbit torque layer. In one embodiment, the spin-orbit torque layer, the magnetic layer, and the metal oxide layer are patterned such that the spin-orbit torque layer is cross-shaped in plan view, and the magnetic layer and the metal oxide layer are disposed over a central portion of the cross-shaped spin-orbit torque layer.

[0060] Another embodiment of the disclosure is a magnetic random access memory (MRAM)-based non-volatile device with physically unclonable functionality (PUF) comprising a spin-orbit torque layer having a first region extending in a first direction over a substrate. A plurality of second regions of the spin-orbit torque layer extend along a second direction perpendicular to the first direction. The plurality of second regions are spaced apart from each other along the first direction. A plurality of first magnetic layers are disposed on the first region of the spin-orbit torque layer. The first magnetic layers are each located between a pair of second regions extending in the second direction from the first region of the spin-orbit torque layer. A metal oxide layer is disposed over each of the first magnetic layers.In one embodiment, the device includes a second magnetic layer disposed over each of the metal oxide layers. In one embodiment, the device includes an intermediate layer disposed over each of the metal oxide layers and the second magnetic layers. In one embodiment, the device includes a conductive layer disposed over each of the metal oxide layers. In one embodiment, the spin-orbit torque layer comprises one or more of platinum, tungsten, tantalum, and PtMn. In one embodiment, the substrate includes a semiconductor substrate with an insulating layer interposed between the semiconductor substrate and the spin-orbit torque layer. In one embodiment, the first magnetic layer comprises a CoFeB material. In one embodiment, the first magnetic layers have a coercivity greater than about 500 gauss.In one embodiment, a magnetic alignment of one or more of the first magnetic layers is achieved by applying a current density of 10. 11 A / m2 to the spin-orbit torque layer.

[0061] Another embodiment of the disclosure is a method of fabricating a magnetic random access memory-based non-volatile Hall effect device, comprising forming a spin-orbit torque layer over a substrate and patterning the spin-orbit torque layer to form a cross-shaped spin-orbit torque layer over the substrate. A first magnetic layer is formed over a central portion of the cross-shaped spin-orbit torque layer, and a metal oxide layer is formed over the first magnetic layer. In one embodiment, in plan view, portions of the spin-orbit torque layer extend outward from the first magnetic layer and the metal oxide layer on opposite sides of a first direction and opposite sides of a second direction, the second direction being perpendicular to the first direction.In one embodiment, the method comprises forming a mask layer over the spin-orbit torque layer prior to forming the first magnetic layer. In one embodiment, the method comprises forming an opening in the mask layer exposing the central portion of the spin-orbit torque layer. In one embodiment, the mask layer comprises a photoresist layer. In one embodiment, the metal oxide layer comprises a material that maintains a perpendicular magnetic anisotropy of the first magnetic layer. In one embodiment, the method comprises forming a conductive layer over the first magnetic layer. In one embodiment, the method comprises forming a second magnetic layer over the first magnetic layer.In one embodiment, the method comprises forming a mask having an opening over a central portion of the spin-orbit torque layer, wherein the second magnetic layer is formed in the mask opening. In one embodiment, the method comprises forming an intermediate layer between the first magnetic layer and the second magnetic layer. In one embodiment, the method comprises forming a mask having an opening over a central portion of the spin-orbit torque layer, wherein the intermediate layer is formed in the mask opening.

[0062] Another embodiment of the disclosure is a method of fabricating a magnetic random access memory-based device, comprising forming a spin-orbit torque layer over a substrate and forming a first magnetic layer over the spin-orbit torque layer. A metal oxide layer is formed over the magnetic layer, and a second magnetic layer is formed over the metal oxide layer. The spin-orbit torque layer, the first magnetic layer, the metal oxide layer, and the second magnetic layer are structured such that, in plan view, portions of the spin-orbit torque layer extend outward from the first magnetic layer, the metal oxide layer, and the second magnetic layer on opposite sides of a first direction and opposite sides of a second direction, the second direction being perpendicular to the first direction.In one embodiment, the method comprises forming a conductive layer over the second magnetic layer. In one embodiment, the method comprises forming an intermediate layer between the metal oxide layer and the second magnetic layer. In one embodiment, the metal oxide layer comprises a material that maintains a perpendicular magnetic anisotropy of the first magnetic layer. In one embodiment, the first magnetic layer has a coercivity greater than about 500 gauss. In one embodiment, the magnetic layer comprises a CoFeB material. In one embodiment, the spin-orbit torque layer comprises one or more of platinum, tungsten, tantalum, and PtMn. In one embodiment, the substrate comprises a semiconductor substrate. In one embodiment, the substrate comprises silicon.In one embodiment, the substrate comprises a semiconductor substrate with an insulating layer interposed between the semiconductor substrate and the spin-orbit torque layer. In one embodiment, the spin-orbit torque layer, the magnetic layer, and the metal oxide layer are patterned such that the spin-orbit torque layer is cross-shaped in plan view, and the magnetic layer and the metal oxide layer are disposed over a central portion of the cross-shaped spin-orbit torque layer.

[0063] Another embodiment is a method of fabricating a magnetic random access memory-based device comprising forming a spin-orbit torque layer over a substrate and forming a first magnetic layer over the spin-orbit torque layer. A metal oxide layer is formed over the magnetic layer. A second magnetic layer is formed over the metal oxide layer. The spin-orbit torque layer, the first magnetic layer, the metal oxide layer, and the second magnetic layer are patterned such that, in plan view, portions of the spin-orbit torque layer extend outward from the first magnetic layer, the metal oxide layer, and the second magnetic layer on opposite sides of a first direction and opposite sides of a second direction, the second direction being perpendicular to the first direction. In one embodiment, the method comprises forming a conductive layer over the second magnetic layer.In one embodiment, the method comprises forming an intermediate layer between the metal oxide layer and the second magnetic layer. In one embodiment, the metal oxide layer comprises a material that maintains a perpendicular magnetic anisotropy of the first magnetic layer. In one embodiment, the first magnetic layer has a coercivity greater than about 500 gauss. In one embodiment, the magnetic layer comprises a CoFeB material. In one embodiment, the spin-orbit torque layer comprises one or more of platinum, tungsten, tantalum, and PtMn. In one embodiment, the substrate comprises a semiconductor substrate. In one embodiment, the substrate comprises silicon. In one embodiment, the substrate comprises a semiconductor substrate with an insulating layer interposed between the semiconductor substrate and the spin-orbit torque layer.In one embodiment, the spin-orbit torque layer, the magnetic layer, and the metal oxide layer are patterned such that the spin-orbit torque layer is cross-shaped in plan view, and the magnetic layer and the metal oxide layer are disposed over a central portion of the cross-shaped spin-orbit torque layer.

[0064] Another embodiment is a magnetic random access memory-based device comprising forming a spin-orbit torque layer disposed over a substrate and a first magnetic layer disposed over the spin-orbit torque layer. A metal oxide layer is disposed over the magnetic layer, and a second magnetic layer is disposed over the metal oxide layer. Portions of the spin-orbit torque layer extend outwardly, in plan view, from the first magnetic layer, the metal oxide layer, and the second magnetic layer on opposite sides of a first direction and opposite sides of a second direction, the second direction being perpendicular to the first direction. In one embodiment, the metal oxide layer comprises a material that maintains a perpendicular magnetic anisotropy of the first magnetic layer.In one embodiment, the first magnetic layer has a coercivity greater than about 500 gauss. In one embodiment, the metal oxide layer comprises MgO. In one embodiment, the first magnetic layer comprises a CoFeB material. In one embodiment, the spin-orbit torque layer comprises one or more of platinum, tungsten, tantalum, and PtMn. In one embodiment, the substrate comprises a semiconductor substrate. In one embodiment, the substrate comprises a semiconductor substrate with an insulating layer interposed between the semiconductor substrate and the spin-orbit torque layer. In one embodiment, the spin-orbit torque layer is cross-shaped in plan view, and the magnetic layer and the metal oxide layers are disposed over a central portion of the cross-shaped spin-orbit torque layer. In one embodiment, the device comprises a conductive layer disposed over the second magnetic layer.In one embodiment, the method includes an intermediate layer disposed between the metal oxide layer and the second magnetic layer. In one embodiment, magnetic alignment of the first magnetic layer is achieved by applying a current density of 10. 11 A / m2 to the spin-orbit torque layer. In one embodiment, the device maintains its magnetic orientation for 10 years without the application of an external current or magnetic field.

[0065] Another embodiment is a method of writing to a magnetic random access memory (MRAM)-based device, comprising determining a magnetic orientation of a first magnetic layer of the MRAM-based non-volatile Hall effect device. The device includes a spin-orbit torque layer disposed over a substrate, the first magnetic layer disposed over the spin-orbit torque layer, and a metal oxide layer disposed over the first magnetic layer. A current density of at least 10 11 A / m2 at a spin-orbit torque layer. In one embodiment, the method comprises determining whether the magnetic orientation of the first magnetic layer changes after applying a current density of at least 10 11A / m2 to a spin-orbit torque layer. In one embodiment, determining whether the magnetic orientation of the first magnetic layer has changed comprises applying a read current to the MRAM-based device. In one embodiment, determining whether the magnetic orientation of a first magnetic layer has changed comprises applying a read current to the MRAM-based device.

[0066] Another embodiment is a method for generating random data, including initializing a plurality of magnetic random access memory-based cells such that the magnetic layer of each of the plurality of magnetic random access memory-based cells is aligned in a first orientation. Each of the plurality of magnetic random access memory-based cells includes a spin-orbit torque layer, the magnetic layer disposed over the spin-orbit torque layer, and a non-magnetic layer disposed over the magnetic layer.Portions of the spin-orbit torque layer extend outwardly in plan view from the magnetic layer and the non-magnetic layer on opposite sides of a first direction and opposite sides of a second direction, the second direction being perpendicular to the first direction, and adjacent magnetic random access memory-based cells are connected to each other along the first direction by the spin-orbit torque layer. A current is applied to the spin-orbit torque layer along the first direction such that the magnetic orientation of the magnetic layer of one or more of the magnetic random access memory-based cells is changed from the first orientation to a second orientation. In one embodiment, the method comprises varying the current to vary a number of magnetic layers that transition from the first orientation to the second orientation.In one embodiment, initializing a plurality of magnetic random access memory-based cells such that the magnetic layer of each of the plurality of magnetic random access memory-based cells is oriented in a first direction by applying an external magnetic field to the plurality of magnetic random access memory-based cells. In one embodiment, initializing a plurality of magnetic random access memory-based cells such that a magnetic layer of each of the plurality of magnetic random access memory-based cells is oriented in a first direction by applying current in the first direction to the spin-orbit torque layer. In one embodiment, the plurality of magnetic random access memory-based cells is a linear array. In one embodiment, the plurality of magnetic random access memory-based cells is a two-dimensional array in plan view.In one embodiment, the non-magnetic layer is a metal oxide layer. In one embodiment, the metal oxide is MgO. In one embodiment, each of the magnetic random access memory-based cells has a second magnetic layer disposed over the non-magnetic layer.

[0067] It will be understood that not all advantages have necessarily been discussed here, no particular advantage is required for all embodiments or examples, and different embodiments or examples may provide different advantages.

Claims

[1] A magnetic random access memory-based non-volatile Hall effect device comprising: a spin-orbit torque layer (10) disposed over a substrate (5); a magnetic layer (20) disposed over the spin-orbit torque layer (10); and a metal oxide layer (30) arranged over the magnetic layer (20), wherein portions of the spin-orbit torque layer (10) extend outwardly from the magnetic layer (20) and the metal oxide layer (30) on opposite sides of a first direction and opposite sides of a second direction in a plan view, the second direction being perpendicular to the first direction, wherein the spin-orbit torque layer (10) comprises heavy metals or materials doped with heavy metals or PtMn, wherein the heavy metals are selected from A and the heavy metal-doped materials are selected from M doped with B; wherein A comprises: Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Cd, In, Hf, Re, Os, Ir, Au, Hg, Tl, Pb, Po, At and / or combinations thereof; where M comprises at least one of the following: Al, Ti, Si, GaMn, GaAs; and wherein B comprises at least one of the following: Fe, Co, Ni, P, S, Ge, As, Se, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Cd, In, Sb, Te, I, Lu, Os, Ir, Bi, Po, At, La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb. [2] The magnetic random access memory-based non-volatile Hall effect device of claim 1, wherein the metal oxide layer (30) comprises a material that maintains a perpendicular magnetic anisotropy of the magnetic layer (20). [3] A magnetic random access memory-based non-volatile Hall effect device according to claim 1 or 2, wherein the magnetic layer (20) has a coercive field strength of greater than 500 x 10 -4 Tesla has. [4] A magnetic random access memory-based non-volatile Hall effect device according to any preceding claim, wherein the metal oxide layer (30) comprises MgO. [5] A magnetic random access memory-based non-volatile Hall effect device according to any preceding claim, wherein the magnetic layer (20) comprises a CoFeB material. [6] A magnetic random access memory-based non-volatile Hall effect device according to any preceding claim, wherein the substrate (5) comprises a semiconductor substrate. [7] A magnetic random access memory-based non-volatile Hall effect device according to any preceding claim, wherein the substrate (5) comprises a semiconductor substrate having an insulating layer interposed between the semiconductor substrate and the spin-orbit torque layer (10). [8] A magnetic random access memory-based non-volatile Hall effect device according to any preceding claim, wherein the spin-orbit torque layer (10) is cross-shaped in a plan view, and the magnetic layer (20) and the metal oxide layer (30) are disposed over a central portion of the cross-shaped spin-orbit torque layer (10). [9] A magnetic random access memory-based non-volatile Hall effect device according to any preceding claim, further comprising a second magnetic layer disposed over the metal oxide layer (30). [10] Magnetic random access memory, MRAM-based non-volatile Hall effect device with physically unclonable function, PUF, comprising: a spin-orbit torque layer (10) having a first region extending in a first direction disposed over a substrate (5); a plurality of second regions of the spin-orbit torque layer (10) extending along a second direction perpendicular to the first direction, wherein the plurality of second regions are spaced apart from each other along the first direction; a plurality of first magnetic layers (20) arranged on the first region of the spin-orbit torque layer (10), wherein the first magnetic layers (20) are each arranged between a pair of second regions extending in the second direction from the first region of the spin-orbit torque layer (10); and a metal oxide layer (30) disposed over each of the first magnetic layers (20), wherein the spin-orbit torque layer (10) comprises heavy metals or materials doped with heavy metals or PtMn, wherein the heavy metals are selected from A and the heavy metal-doped materials are selected from M doped with B; wherein A comprises: Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Cd, In, Hf, Re, Os, Ir, Au, Hg, Tl, Pb, Po, At and / or combinations thereof; where M comprises at least one of the following: Al, Ti, Si, GaMn, GaAs; and wherein B comprises at least one of the following: Fe, Co, Ni, P, S, Ge, As, Se, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Cd, In, Sb, Te, I, Lu, Os, Ir, Bi, Po, At, La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb. [11] The MRAM-based non-volatile PUF device of claim 10, further comprising a second magnetic layer disposed over each of the metal oxide layers (30). [12] The MRAM-based non-volatile PUF device of claim 11, further comprising an intermediate layer disposed between each of said metal oxide layers (30) and second magnetic layers. [13] The MRAM-based non-volatile PUF device of any one of claims 10 to 12, further comprising a conductive layer disposed over each of the metal oxide layers (30). [14] The MRAM-based non-volatile PUF device according to any one of the preceding claims 10 to 13, wherein the substrate (5) comprises a semiconductor substrate having an insulating layer interposed between the semiconductor substrate and the spin-orbit torque layer (10). [15] The MRAM-based PUF device of any one of claims 10 to 14, wherein the first magnetic layer (20) comprises a CoFeB material. [16] MRAM-based PUF device according to one of the preceding claims 10 to 15, wherein the first magnetic layer (20) has a coercive field strength of greater than 500 x 10 -4 Tesla has. [17] A method of manufacturing a magnetic random access memory-based non-volatile Hall effect device, comprising: Forming a spin-orbit torque layer (10) over a substrate (5); Structuring the spin-orbit torque layer (10) to form a cross-shaped spin-orbit torque layer (10) over the substrate (5); forming a first magnetic layer (20) over a central portion of the cross-shaped spin-orbit torque layer (10); and Forming a metal oxide layer (30) over the first magnetic layer (20), wherein the spin-orbit torque layer (10) comprises heavy metals or materials doped with heavy metals or PtMn, wherein the heavy metals are selected from A and the heavy metal-doped materials are selected from M doped with B; wherein A comprises: Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Cd, In, Hf, Re, Os, Ir, Au, Hg, Tl, Pb, Po, At and / or combinations thereof; where M comprises at least one of the following: Al, Ti, Si, GaMn, GaAs; and wherein B comprises at least one of the following: Fe, Co, Ni, P, S, Ge, As, Se, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Cd, In, Sb, Te, I, Lu, Os, Ir, Bi, Po, At, La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb. [18] The method of claim 17, wherein portions of the spin-orbit torque layer (10) extend outwardly in plan view from the first magnetic layer (20) and the metal oxide layer (30) on opposite sides of a first direction and opposite sides of a second direction, the second direction being perpendicular to the first direction.

Images