Spin-orbit torque random number generator and its manufacturing method

Abstract

To provide: a spin orbital torque random number generator that implements a random number generator having a low switching current value and a high-entropy random number generation function; and a manufacturing method thereof.SOLUTION: A spin orbital torque random number generator 100 includes: a spin torque generation layer including a first spin torque layer and a second spin torque layer; a magnetization free layer; a tunnel barrier layer; and a magnetization fixed layer. The second spin torque layer is positioned between the magnetization free layer and the first spin torque layer, and generates: a field torque related to a magnetization direction of the magnetization free layer; and a damping torque. The first spin torque layer is associated with a thickness of the second spin torque layer, and adjusts a ratio of a torque according to a strength of the field torque to a strength of the damping torque. The magnetization free layer randomly aligns, on the basis of the adjusted torque ratio, the magnetization direction in an upper or lower direction, and generates, on the basis of the randomly aligned magnetization direction, a random number based on a current that is switched and output.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of priority based on Korean Patent Application No. 10-2023-0153395, filed November 8, 2023, and all contents disclosed in the documents of this Korean patent application are incorporated herein by reference.

[0002] The present invention relates to a spin-orbit torque random number generator and a manufacturing method thereof, and more particularly to a technology for realizing a random number generator capable of switching spin-orbit torque based on a spin torque layer, and having a low switching current value and a high entropy random number generation function by controlling the ratio of damping-like torque (DLT) and field-like torque (FLT) components of the spin-orbit torque. [Background technology]

[0003] A magnetic tunnel junction (MTJ), which is the core structure of a spin-orbit torque (SOT) switching-based device, can be composed of a non-magnetic spin torque generating layer (hereinafter referred to as the "spin torque layer"), a first magnetic layer (a magnetization free layer, hereinafter referred to as the "free layer"), a tunnel barrier layer, and a second magnetic layer (a magnetization fixed layer, hereinafter referred to as the "fixed layer").

[0004] Information is read using the tunneling magnetoresistance (TMR) phenomenon, in which the electrical resistance of the tunnel current passing through the insulating layer changes depending on the relative magnetization directions of the free layer and the fixed layer.

[0005] To achieve a high tunnel magnetoresistance ratio, a low write current, and high integration, the magnetic tunnel junction must have perpendicular magnetic anisotropy (PMA) characteristics.

[0006] Perpendicular magnetic anisotropy means that the magnetization direction of the magnetic layer is perpendicular to the plane of the magnetic layer.

[0007] Recently, the spin-orbit torque (SOT) phenomenon has been discovered, which induces switching in the free layer using the spin Hall effect or Rashba effect that occurs when current flows in a direction parallel to the plane of the spin torque layer adjacent to the free layer.

[0008] It is attracting attention as a technology that enables faster switching than the existing spin-transfer torque (STT) writing method.

[0009] At this time, the spin-orbit torque induced in the magnetization free layer by the in-plane current induces magnetization in the in-plane direction.

[0010] When the current supply is interrupted, the magnetization, which was aligned in the in-plane direction, becomes randomly aligned in the up or down direction perpendicular to the magnetic layer surface. Based on this random alignment, it can be used as a true random number generator (TRNG).

[0011] A random number generator (RNG) is a device that generates a sequence of numbers or symbols that are theoretically unpredictable.

[0012] The random number generation function is a standard for systematically creating and applying arbitrary numbers, and can be a function for generating numbers that can be applied to encryption techniques.

[0013] The key to realizing a spin-orbit torque random number generator is to determine what material and structure to use in the spin torque layer to maintain high entropy of the random number sequence and achieve low write current.

[0014] Prior art (Appl. Phys. Lett. 118, 052401 (2021)) illustrates a Ta / CoFeB / MgO / Ta-based spin-orbit torque random number generator for generating random numbers in artificial neural networks.

[0015] A system is designed that has similar interference accuracy and low error rate as existing schemes. Summary of the Invention [Problem to be solved by the invention]

[0016] The present invention aims to realize a spin-orbit torque random number generator that is capable of switching spin-orbit torque (SOT) based on a spin torque layer, and has a low switching current value and a high entropy random number generation function by controlling the ratio of damping-like torque (DLT) and field-like torque (FLT) among the components of the spin-orbit torque.

[0017] The present invention aims to realize a spin-orbit torque random number generator that maintains perpendicular magnetic anisotropy and high entropy and has a low switching current by stacking materials having different spin-orbit coupling magnitudes in a spin torque layer that is in contact with a free layer and provides an in-plane current.

[0018] The present invention aims to provide a candidate material for a spin torque layer having low DLT and high FLT efficiency as a material for a SOT for a random number generator, which has an increased ratio of FLT to DLT.

[0019] The present invention aims to realize a spin-orbit torque random number generator with low switching current while maintaining a high entropy random number generation function by adjusting the ratio of FLT to DLT based on the structure of a spin torque layer and a magnetic layer based on a material candidate that increases the ratio of FLT compared to DLT. [Means for solving the problem]

[0020] A spin-orbit torque random number generator according to an embodiment of the present invention includes a spin torque generation layer including a first spin torque layer and a second spin torque layer, a magnetic free layer, a tunnel barrier layer, and a magnetic fixed layer, wherein the second spin torque layer is located between the magnetic free layer and the first spin torque layer and generates a field-like torque (FLT) and a damping-like torque (DLT) associated with the magnetization direction of the magnetic free layer, the first spin torque layer adjusts a torque ratio due to the magnitude of the field torque to the magnitude of the damping torque in cooperation with a thickness of the second spin torque layer, and the magnetic free layer has its magnetization direction randomly aligned in an up or down direction based on the adjusted torque ratio, and can generate a random number based on a current that is switched and output based on the randomly aligned magnetization direction.

[0021] The torque ratio may increase as the magnitude of the damping torque decreases while the magnitude of the field torque is maintained when the thickness of the second spin torque layer increases, and may decrease as the magnitude of the damping torque increases while the magnitude of the field torque is maintained when the thickness of the second spin torque layer decreases.

[0022] The first spin torque layer may be made of any one of Ta, W, and Pt, and the second spin torque layer may be made of Nb.

[0023] The second spin torque layer may be formed to a thickness of 1 nm to 15 nm.

[0024] Based on the first spin torque layer and the second spin torque layer, the magnitude of the switching current that switches the magnetization direction can be reduced and the switching efficiency of switching the magnetization direction can be increased depending on the torque ratio in which the ratio of the field torque among the components of the spin-orbit torque is higher than the damping torque.

[0025] A spin-orbit torque random number generator according to an embodiment of the present invention can realize a random number generation function by having an entropy value of a random number sequence for random numbers based on a current output according to a ratio of the torque to a repeatedly injected current that is higher than a reference value related to randomness.

[0026] A method for manufacturing a spin-orbit torque random number generator according to an embodiment of the present invention includes forming a spin torque generation layer including a first spin torque layer and a second spin torque layer, forming a magnetic free layer on the spin torque generation layer, forming a tunnel barrier layer on the magnetic free layer, and forming a magnetic fixed layer on the tunnel barrier layer to form a spin-orbit torque random number generator, wherein the second spin torque layer is located between the magnetic free layer and the first spin torque layer and generates a field-like torque (FLT) and a damping-like torque (DLT) associated with a magnetization direction of the magnetic free layer, the first spin torque layer adjusts a torque ratio according to a magnitude of the field torque to a magnitude of the damping torque in cooperation with a thickness of the second spin torque layer, and the magnetization direction of the magnetic free layer is randomly aligned in an up or down direction based on the adjusted torque ratio, and the spin-orbit torque random number generator can generate random numbers based on a current that is switched and output based on the randomly aligned magnetization direction.

[0027] The torque ratio may increase as the magnitude of the damping torque decreases while the magnitude of the field torque is maintained when the thickness of the second spin torque layer increases, and may decrease as the magnitude of the damping torque increases while the magnitude of the field torque is maintained when the thickness of the second spin torque layer decreases.

[0028] The step of forming the spin torque generation layer can include the step of forming the first spin torque layer from any one of Ta, W, and Pt, and the step of forming the second spin torque layer from Nb.

[0029] The second spin torque layer may be formed to a thickness of 1 nm to 15 nm. [Effects of the Invention]

[0030] The present invention can realize a spin-orbit torque random number generator that is capable of switching spin-orbit torque (SOT) based on a spin torque layer, and has a low switching current value and a high entropy random number generation function by controlling the ratio of damping-like torque (DLT) and field-like torque (FLT) components of the spin-orbit torque.

[0031] The present invention can realize a spin-orbit torque random number generator that maintains perpendicular magnetic anisotropy and high entropy and has a low switching current by stacking materials having different spin-orbit coupling magnitudes in a spin torque layer that contacts a free layer and provides an in-plane current.

[0032] The present invention can provide a candidate material for a spin torque layer having low DLT and high FLT efficiency as a material for a SOT for a random number generator, in which the ratio of FLT to DLT is increased.

[0033] The present invention is based on the structure of a spin torque layer and a magnetic layer based on a material candidate that increases the ratio of FLT to DLT, and by adjusting the ratio of FLT to DLT, it is possible to realize a spin-orbit torque random number generator with low switching current while maintaining the function of generating random numbers with high entropy. [Brief explanation of the drawings]

[0034] [Figure 1] 1A and 1B are diagrams illustrating the structure and configuration of a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 2] 10A and 10B are diagrams illustrating the random number generation operation of a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 3] FIG. 10 is a diagram illustrating a control structure of the ratio between damping torque and field torque in a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 4A] FIG. 2 is a diagram illustrating a three-dimensional structure of a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 4B] FIG. 2 is a diagram illustrating a three-dimensional structure of a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 4C] FIG. 2 is a diagram illustrating a three-dimensional structure of a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 5A] FIG. 10 is a diagram illustrating perpendicular magnetic anisotropy of a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 5B] FIG. 10 is a diagram illustrating perpendicular magnetic anisotropy of a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 6] 10A and 10B are diagrams illustrating simulation results for controlling the ratio of damping torque to field torque in a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 7] 10A and 10B are diagrams illustrating simulation results for controlling the ratio of damping torque to field torque in a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 8A]10A and 10B are diagrams illustrating switching operations and efficiency associated with controlling the ratio between damping torque and field torque in a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 8B] 10A and 10B are diagrams illustrating switching operations and efficiency associated with controlling the ratio between damping torque and field torque in a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 9] 10A and 10B are diagrams illustrating switching operations and efficiency associated with controlling the ratio between damping torque and field torque in a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 10] 10A and 10B are diagrams illustrating switching operations and efficiency associated with controlling the ratio between damping torque and field torque in a spin-orbit torque random number generator according to an embodiment of the present invention. [Figure 11] 10A and 10B are diagrams illustrating the results of a switching probability test and an entropy test of a random number sequence of a spin-orbit torque random number generator according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0035] Various embodiments of the present document will now be described with reference to the accompanying drawings.

[0036] The examples and the terms used therein are not intended to limit the technology described in this document to a particular embodiment, but should be understood to include various modifications, equivalents, and / or alternatives to the examples.

[0037] In the following description of various embodiments, if it is determined that a detailed description of related publicly known functions or configurations may obscure the gist of the invention, the detailed description will be omitted.

[0038] The terms used below are defined in consideration of the functions in various embodiments, and may vary depending on the intentions or practices of users or operators. Therefore, the definitions should be based on the overall content of this specification.

[0039] In connection with the description of the drawings, like reference numerals may be used for like components.

[0040] A singular expression can include a plural expression unless the context clearly indicates otherwise.

[0041] In this document, phrases such as "A or B" or "at least one of A and / or B" may include all possible combinations of the items listed together.

[0042] Expressions such as "first," "second," "initial," or "second" can modify the components in question regardless of order or importance, and are used only to distinguish one component from other components, and do not limit the components in question.

[0043] When a (e.g., first) component is referred to as being "(functionally or communicatively) coupled" or "connected" to another (e.g., second) component, the component may be directly coupled to the other component or may be coupled through another component (e.g., third component).

[0044] As used herein, "configured to" may be used interchangeably with, for example, hardware or software "suitable for," "capable of," "modified to," "made to," "capable of," or "designed to," depending on the context.

[0045] In some circumstances, the phrase "a device configured to" can mean that the device is "capable of" in conjunction with other devices or components.

[0046] For example, the phrase "a processor configured (or set) to perform A, B, and C" may refer to a processor that is dedicated to performing those operations (e.g., an embedded processor), or to a general-purpose processor (e.g., a CPU or application processor) that can perform those operations by executing one or more software programs stored in a memory device.

[0047] Also, the term "or" means an inclusive or rather than an exclusive or.

[0048] That is, unless otherwise stated or clear from the context, the phrase "x uses a or b" means any one of the natural inclusive permutations.

[0049] As used below, terms such as "...module" and "...device" refer to a unit that processes at least one function or operation, which may be embodied in hardware, software, or a combination of hardware and software.

[0050] FIG. 1 is a diagram illustrating the structure and configuration of a spin-orbit torque random number generator according to an embodiment of the present invention.

[0051] FIG. 1 illustrates the structure and configuration of a spin-orbit torque random number generator according to one embodiment of the present invention.

[0052] Referring to FIG. 1, a spin-orbit torque random number generator 100 according to one embodiment of the present invention includes a first spin torque layer 110, a second spin torque layer 111, a magnetic free layer 120, and a magnetic fixed layer 121, and includes a tunnel barrier layer 130 between the magnetic free layer 120 and the magnetic fixed layer 121.

[0053] For example, the spin-orbit torque random number generator 100 may be referred to as a true random number generator (TRNG).

[0054] As an example, the first spin torque layer 110 and the second spin torque layer 111 may be configured as spin torque generation layers.

[0055] A manufacturing method of the spin-orbit torque random number generator 100 according to one embodiment of the present invention may include the steps of forming a spin torque generation layer including a first spin torque layer 110 and a second spin torque layer 111, forming a magnetization free layer 120 on the spin torque generation layer, forming a tunnel barrier layer 130 on the magnetization free layer 120, and forming a magnetization fixed layer 121 on the tunnel barrier layer 130 to form the spin-orbit torque random number generator 100.

[0056] As an example, the second spin torque layer 111 is located between the magnetization free layer 120 and the first spin torque layer 110 and can generate a field-like torque (FLT) and a damping-like torque (DLT) associated with the magnetization direction of the magnetization free layer 120.

[0057] The first spin torque layer 110 can adjust the ratio of the magnitude of the field torque to the magnitude of the damping torque in cooperation with the thickness of the second spin torque layer 111.

[0058] The magnetization direction of the free magnetic layer 120 is randomly aligned in the up or down direction based on the ratio of the adjusted torques.

[0059] This allows the spin-orbit torque random number generator 100 to generate random numbers based on currents that are switched on and off based on randomly aligned magnetization directions.

[0060] Depending on whether the randomly aligned magnetization direction of the magnetization free layer 120 is the same as or opposite to the magnetization direction of the magnetization fixed layer 121, the resistance state of the spin orbit torque random number generator 100 is determined to be a high impedance state or a low impedance state, and the spin orbit torque random number generator 100 can generate random numbers by determining an arbitrary number based on the current output according to the determined state.

[0061] For example, when the thickness of the second spin torque layer 111 increases, the torque ratio increases as the magnitude of the damping torque decreases while the magnitude of the field torque is maintained, and when the thickness of the second spin torque layer 111 decreases, the torque ratio decreases as the magnitude of the damping torque increases while the magnitude of the field torque is maintained.

[0062] The first spin torque layer 110 may be formed of any one of Ta, W, and Pt.

[0063] The second spin torque layer 111 may be made of Nb.

[0064] The first spin torque layer 110 is formed of a metal having high damping torque efficiency, and the second spin torque layer 111 is formed of a metal having high field torque efficiency, but may be formed of a material having lower damping torque efficiency than the first spin torque layer 110.

[0065] That is, the materials forming the first spin torque layer 110 and the second spin torque layer 111 are not limited to the above examples, and may be replaced with materials that can achieve the same performance.

[0066] For example, the second spin torque layer 111 may be formed to a thickness of 1 nm to 15 nm.

[0067] As an example, the spin-orbit torque random number generator 100 can reduce the magnitude of the switching current that switches the magnetization direction and increase the switching efficiency of switching the magnetization direction based on the first spin torque layer 110 and the second spin torque layer 111, depending on the ratio of the field torque component of the spin-orbit torque that is higher than the damping torque component.

[0068] The spin-orbit torque random number generator 100 according to an embodiment of the present invention can realize a random number generation function by having an entropy value of a random number sequence for random numbers based on a current output according to a ratio of the torque to the repeatedly injected current that is higher than a reference value related to randomness.

[0069] Therefore, the present invention can realize a spin-orbit torque random number generator that is capable of switching spin-orbit torque (SOT) based on a spin torque layer, and has a low switching current value and a high entropy random number generation function by controlling the ratio of damping-like torque (DLT) and field-like torque (FLT) components of the spin-orbit torque.

[0070] FIG. 2 is a diagram illustrating the random number generation operation of the spin-orbit torque random number generator according to one embodiment of the present invention.

[0071] FIG. 2 illustrates the principle by which a spin-orbit torque element can operate as a random number generator in connection with the random number generation operation of a spin-orbit torque random number generator according to one embodiment of the present invention.

[0072] Referring to FIG. 2, a first state 200 illustrates the state of the spin-orbit torque element when current is applied, that is, when it is on, and a second state 210 illustrates the state of the spin-orbit torque element when no current is applied, that is, when it is off.

[0073] In the first state 200, when an in-plane current is applied to the spin torque generating layer, the spins are classified into upper and lower spins, and two types of torques are generated by the spin current: damping torque and field torque.

[0074] As can be seen from the block diagram, the magnetization of the magnetization free layer is initially perpendicular due to the two types of torque, and when torque is applied, it is determined to be in the in-plane direction, which is the X direction.

[0075] In other words, when a current is injected, the magnetization, which was stable and perpendicular, lies in the X direction and becomes in-plane.

[0076] On the other hand, when we look at the energy bands, we can see that the energy states are fixed at specific positions.

[0077] The second state 210, where current injection is discontinued, indicates that when the applied in-plane current is discontinued, the magnetization direction drops from the unstable in-plane direction to an equally stable up or down direction.

[0078] On the other hand, when we look at the energy bands, the energy state is not fixed to a specific position, but changes randomly in one of two directions.

[0079] Here, by determining "1" or "0" based on the magnetization direction being changed to the up or down direction, it can operate as a random number generator based on binary data determined randomly.

[0080] On the other hand, the ratio determined to be "1" or "0" and the efficiency of switching are affected by the ratio of field torque to damping torque, but since the weight of field torque is higher than that of damping torque, it can be applied as a random number generation function with low switching current value and high entropy.

[0081] FIG. 3 is a diagram illustrating a control structure of the ratio between the damping torque and the field torque in a spin-orbit torque random number generator according to an embodiment of the present invention.

[0082] FIG. 3 illustrates the ratio of damping torque to field torque depending on the thickness of the second spin torque layer in relation to a control structure for the ratio of damping torque to field torque of a spin-orbit torque random number generator according to one embodiment of the present invention.

[0083] Referring to FIG. 3, a first structure 300 of a spin-orbit torque random number generator according to an embodiment of the present invention includes a first spin torque layer 301, a second spin torque layer 302 and a magnetic free layer 303.

[0084] The second structure 310 of the spin-orbit torque random number generator according to an embodiment of the present invention includes a first spin torque layer 311 , a second spin torque layer 312 and a magnetization free layer 313 .

[0085] When comparing the torque ratio of the field torque 305 to the damping torque 304 of the first structure 300 with the torque ratio of the field torque 315 to the damping torque 314 of the second structure 310, the second structure 310 has an even smaller torque ratio.

[0086] Comparing the thickness of the second spin torque layer 302 with the thickness of the second spin torque layer 312, the thickness of the second spin torque layer 312 is greater and the torque ratio is smaller.

[0087] As a result, the spin-orbit torque random number generator according to one embodiment of the present invention can control the torque ratio by adjusting the thickness of the second spin torque layer, and based on the control of the torque ratio, can maintain perpendicular magnetic anisotropy and high entropy and have a low switching current.

[0088] Therefore, the present invention can realize a spin-orbit torque random number generator that maintains perpendicular magnetic anisotropy and high entropy and has a low switching current by stacking materials having different spin-orbit coupling magnitudes in a spin torque layer that contacts the free layer and provides an in-plane current.

[0089] 4A to 4C are diagrams illustrating a three-dimensional structure of a spin-orbit torque random number generator according to an embodiment of the present invention.

[0090] 4A and 4B illustrate a three-dimensional structure of a spin-orbit torque random number generator according to one embodiment of the present invention.

[0091] Referring to FIG. 4A, a three-dimensional structure 400 of a spin-orbit torque random number generator according to an embodiment of the present invention is a structure for measuring the efficiency of spin-orbit torque (SOT), and has a structure in which a magnetization free layer is formed on a spin-torque generating layer.

[0092] Sputtering can be used to deposit the thin films required to fabricate the structure of the elements of the spin-orbit torque random number generator according to one embodiment of the present invention.

[0093] The initial vacuum (base pressure) is 5×10 -9 The deposition is carried out in an argon (Ar) atmosphere at a temperature of less than Torr.

[0094] The thickness of each layer of the spin-orbit torque random number generator can be adjusted using deposition time and sputtering power.

[0095] The three-dimensional structure 400 of the spin-orbit torque random number generator includes a first spin torque layer 401 , a second spin torque layer 402 and a magnetization free layer 403 .

[0096] The first spin torque layer 401 is formed on a substrate, and a native oxide layer may be formed on the substrate.

[0097] The substrate may be formed of Si, and the native oxide layer may be formed of SiO2 and may be amorphous.

[0098] The first spin torque layer 401 and the second spin torque layer 402 can be deposited using DC magnetron sputtering.

[0099] The size of the target used to prepare the thin film may be 2 inches in diameter and the power may be 50W.

[0100] The magnetic free layer 403 may be formed of CoFeB, and the composition of the sputtering target may be Co 40 Fe 40 B 20 (at %).

[0101] Additionally, an insulating layer may be formed of MgO and a capping layer may be formed of Ta.

[0102] After the deposition of the thin film, the film was heat-treated at 300°C for 1 hour. The initial vacuum during the heat treatment was 10 -6 Torr range, and an external magnetic field of 6 kOe can be applied perpendicular to the thin film during the heat treatment.

[0103] The three-dimensional structure 400 of the spin-orbit torque random number generator is 5 × 35 μm 2 The cross-shaped Hall bar may be formed in a size of 1 / 2.

[0104] Referring to FIG. 4B, a three-dimensional structure 410 of a spin-orbit torque random number generator according to an embodiment of the present invention is a structure for measuring the efficiency of spin-orbit torque (SOT), and has a structure in which a magnetic free layer is formed on a spin-torque generating layer.

[0105] The three-dimensional structure 410 of the spin-orbit torque random number generator includes a first spin torque layer 411 , a second spin torque layer 412 and a magnetization free layer 413 .

[0106] The three-dimensional structure 410 of the spin-orbit torque random number generator can be formed in the shape of an island with a diameter of 5 μm to confirm the SOT switching characteristics.

[0107] The three-dimensional structure 410 of the spin-orbit torque random number generator according to an embodiment of the present invention is a structure for measuring the efficiency of spin-orbit torque (SOT), and has a structure in which a magnetization free layer is formed on a spin-torque generating layer.

[0108] FIG. 4C illustrates the generation of random numbers in a three-dimensional structure of a spin-orbit torque random number generator according to one embodiment of the present invention.

[0109] Referring to FIG. 4C, a first state 420 indicates the magnetization direction of the magnetic free layer in which the spin-orbit torque random number generator corresponds to “On”, and a second state 421 indicates an “Off” operation in which the magnetic free layer of the spin-orbit torque random number generator has a magnetization direction corresponding to “1” or “0”.

[0110] During the injection of the current pulse, the magnetization is oriented in-plane, and upon interruption of the current supply, it becomes randomly aligned in the z-direction, which corresponds to the perpendicular direction.

[0111] The alignment direction can be read out as a switching probability distribution corresponding to the lateral Hall voltage.

[0112] 5A and 5B are diagrams illustrating perpendicular magnetic anisotropy of a spin-orbit torque random number generator according to one embodiment of the present invention.

[0113] 5A and 5B illustrate magnetic hysteresis loops measured in a perpendicular direction using a vibrating sample magnetometer (VSM) in relation to perpendicular magnetic anisotropy of a spin-orbit torque random number generator according to an embodiment of the present invention.

[0114] Referring to Figures 5A and 5B, graph 500 shows a first embodiment in which the spin torque generation layer does not have a first spin torque layer, the second spin torque layer is formed of niobium (Nb), and the thickness is adjusted to 5 nm to 15 nm, and graph 501 shows a magnetic hysteresis curve measured by applying a magnetic field in the in-plane direction of the thin film in the structure according to the first embodiment described above.

[0115] Graph 502 shows a second example in which the spin torque generation layer is formed of a first spin torque layer made of tantalum (Ta) with a thickness of approximately 3 nm and a second spin torque layer made of niobium (Nb) with a thickness adjusted to 5 nm to 15 nm, and graph 503 shows a magnetic hysteresis curve measured by applying a magnetic field in the in-plane direction of the thin film in the structure according to the second example described above.

[0116] Graph 504 shows a third embodiment in which the spin torque generation layer is formed of a first spin torque layer made of tungsten (W) with a thickness of approximately 3 nm and a second spin torque layer made of niobium (Nb) with a thickness adjusted to 5 nm to 15 nm, and graph 505 shows a magnetic hysteresis curve measured by applying a magnetic field in the in-plane direction of the thin film in the structure according to the third embodiment described above.

[0117] Graph 506 shows a fourth embodiment in which the spin torque generation layer is formed of a first spin torque layer made of platinum (Pt) with a thickness of approximately 3 nm and a second spin torque layer made of niobium (Nb), with thicknesses adjusted to 1 nm to 15 nm, and graph 507 shows a magnetic hysteresis curve measured by applying a magnetic field in the in-plane direction of the thin film in the structure according to the fourth embodiment described above.

[0118] Graphs 501, 503, 505, and 507 show that when the second spin torque layer is made of niobium, it has perpendicular magnetic anisotropy with respect to the magnetic free layer.

[0119] It can also be confirmed that perpendicular magnetic anisotropy is exhibited even when the thickness of the second spin torque layer is changed.

[0120] 6 and 7 are diagrams illustrating experimental measurement results for controlling the ratio between the damping torque and the field torque in a spin-orbit torque random number generator according to an embodiment of the present invention.

[0121] FIG. 6 illustrates the results of measuring the efficiency of the damping torque (ξ DL ) and field torque (ξ FL ) as a function of the thickness of niobium measured using a harmonic measurement method in Nb / CoFeB / MgO / Ta and Ta(W,Pt) / Nb / CoFeB (thickness is variable in the case of Pt) / MgO / Ta structures in connection with controlling the ratio between the damping torque and field torque in a spin-orbit torque random number generator according to an embodiment of the present invention.

[0122] Referring to FIG. 6, a graph 600 shows measurement results of damping torque and field torque efficiency for the first embodiment, which corresponds to the case where the spin torque generation layer is formed of only Nb.

[0123] Graph 601 shows the measurement results of the damping torque and field torque efficiencies for the second embodiment, in which the spin torque generation layers are made of Ta as the first spin torque layer and Nb as the second spin torque layer.

[0124] Graph 602 shows the measurement results of the damping torque and field torque efficiencies for the third embodiment, which corresponds to the case where the spin torque generation layers are made of a first spin torque layer made of W and a second spin torque layer made of Nb.

[0125] Graph 603 shows the measurement results of the damping torque and field torque efficiencies for the fourth embodiment, which corresponds to the case where the spin torque generation layers are made of Pt as the first spin torque layer and Nb as the second spin torque layer.

[0126] Graph 600 shows that the field torque and damping torque tend to increase with increasing niobium thickness, and then gradually saturate.

[0127] The damping torque efficiency is approximately -0.03, which is lower than that of other heavy metals (W: -0.33, Ta: -0.12, Pt: 0.09, etc.), but the field torque efficiency is 0.09, and the field torque / damping torque ratio (hereinafter referred to as "η") is high at approximately 3.

[0128] In the case of the Ta, W, Pt / Nb structures shown in graphs 601 to 603, heavy metals with high damping torque efficiency are intentionally placed under the niobium layer, so the smaller the thickness of the niobium, the stronger the influence of the damping torque that the Ta, W, and Pt layers exert on the free layer, and the damping torque efficiency of the entire element tends to increase.

[0129] On the other hand, in the case of field torque, which is known to have characteristics originating from the interface between the non-magnetic layer and the magnetic free layer, the rate of change due to the thickness of niobium is small compared to the damping torque.

[0130] Characteristically, in the case of the Pt / Nb structure, the signs of the spin-orbit coupling of Pt and Nb are opposite to each other, so the torques emitted from the two layers compete with each other, and as the influence of niobium increases (as the thickness increases), a point is identified where the sign of the overall efficiency changes, after which it gradually tends to saturate.

[0131] FIG. 7 illustrates the ratio of field torque to damping torque as a function of niobium thickness measured using a harmonic measurement method in Nb / CoFeB / MgO / Ta and Ta(W,Pt) / Nb / CoFeB / MgO / Ta (unit: nm) structures in connection with control of the ratio of damping torque to field torque in a spin-orbit torque random number generator according to an embodiment of the present invention.

[0132] Referring to FIG. 7, graphs 700 and 701 show the ratio of field torque to damping torque (η) for each structure as a function of niobium thickness as identified in FIG.

[0133] The ratio (η) exhibits values ​​varying from −0.5 to −3, and the Nb structure in graph 700 can exhibit the highest ratio (η) among the four structures.

[0134] According to graph 701, the Ta, W, or Pt / Nb structure exhibits high damping torque efficiency because the smaller the influence of niobium, the more it is affected by the large damping torque of the heavy metal layer, whereas the field torque efficiency changes only slightly, and therefore the ratio (η) tends to gradually decrease.

[0135] 8A to 10 are diagrams illustrating switching operations and efficiencies associated with controlling the ratio between damping torque and field torque in a spin-orbit torque random number generator according to an embodiment of the present invention.

[0136] 8A and 8B show the switching probability distributions (P) of eight elements having different ratios (η) in relation to the switching operation and efficiency associated with controlling the ratio between damping torque and field torque in a spin-orbit torque random number generator according to one embodiment of the present invention. SW ), and the resulting switching operating current (I SW ) is shown below.

[0137] Referring to FIG. 8A, graph 800 illustrates the switching probability distribution and switching operation current when the ratio (η) is −1.22, graph 801 illustrates the switching probability distribution and switching operation current when the ratio (η) is −1.53, graph 802 illustrates the switching probability distribution and switching operation current when the ratio (η) is −1.63, and graph 803 illustrates the switching probability distribution and switching operation current when the ratio (η) is −2.02.

[0138] Referring to FIG. 8B, graph 804 illustrates the switching probability distribution and switching operation current when the ratio (η) is −2.18, graph 805 illustrates the switching probability distribution and switching operation current when the ratio (η) is −2.24, graph 806 illustrates the switching probability distribution and switching operation current when the ratio (η) is −2.40, and graph 807 illustrates the switching probability distribution and switching operation current when the ratio (η) is −3.03.

[0139] For example, graphs 800 to 807 are obtained by using a four-point probe to calculate the switching probability distribution (P SW ) and switching operating current (I SW ) indicates the measured results.

[0140] All current pulses were injected in an environment without an external magnetic field, and all pulses had the same length of 10 μsec. Pulses of the same magnitude were injected repeatedly 50 times.

[0141] The SOT due to the small magnitude pulse is not sufficient to completely align the free layer, so the switching probability is zero.

[0142] As the magnitude of the injected pulse increases, the probability of switching occurring increases, eventually converging to 50%.

[0143] This tendency follows a Boltzmann sigmoidal function, which can be used to define the magnitude of the current pulse that results in a switching probability of 49% as the switching current.

[0144] FIG. 9 shows the switching current density (J) of eight elements having different ratios (η) related to the control of the ratio between damping torque and field torque in a spin-orbit torque random number generator according to one embodiment of the present invention. sw ) into the coercive field (H c), the tendency of the value to the ratio (η) is illustrated.

[0145] Referring to FIG. 9, a graph 900 plots the value of the switching current versus the ratio (η) in graphs 800 to 807 described in FIGS. 8A and 8B, as a function of switching current density / coercive field (J sw / H c ), and the value indicates the tendency for the magnitude of the ratio (η).

[0146] The switching current density of a spin-orbit torque device having a multi-domain magnetic layer tends to be proportional to the magnitude of the coercive field.

[0147] Therefore, to see the effect of the ratio (η) alone on the current density of the eight elements with different materials and structures, we can normalize by dividing the coercive field of each element.

[0148] As a result, it can be confirmed that the larger the value of the ratio (η), the higher the switching efficiency.

[0149] Therefore, the present invention can provide a candidate material for a spin torque layer having a low DLT and a high FLT efficiency as a material for a SOT for a random number generator, in which the ratio of FLT to DLT is increased.

[0150] FIG. 10 illustrates the results of a Fokker-Planck calculation using the macrospin approximation showing the effect of the magnitude of the ratio (η) on the switching current density, which is associated with controlling the ratio between the damping torque and the field torque, in a spin-orbit torque random number generator according to one embodiment of the present invention.

[0151] Referring to FIG. 10, graphs 1000 and 1001 are the results of a Fokker-Planck calculation using the macrospin approximation, and show how the magnitude of the ratio (η) affects the switching current density.

[0152] The larger the ratio (η), the lower the current density required to switch the spin-orbit torque device, which confirms the experimental results described in the previous figure.

[0153] In graph 1000, the pulse width of the charging current flowing through the first spin torque layer may be 50 ns and the rise time may be 2 ns.

[0154] Graph 1001 shows the switching current density (J SW ), where the switching probability (P SW ) is 50%.

[0155] Graph 1001 shows that the theoretical and experimental results show similar patterns.

[0156] FIG. 11 is a diagram illustrating the results of a switching probability test and an entropy test of a random number sequence of a spin-orbit torque random number generator according to an embodiment of the present invention.

[0157] FIG. 11 illustrates the results of a switching probability and random number sequence entropy test of a spin-orbit torque random number generator according to one embodiment of the present invention.

[0158] Referring to FIG. 11, graph 1100 shows the switching probability of a spin-orbit torque random number generator including a spin torque generation layer including only the second spin torque layer, and a spin-orbit torque random number generator including a spin torque generation layer including the first and second spin torque layers.

[0159] Three elements with different ratios (η) were selected to generate random number sequences, and the randomness of each string was examined.

[0160] To confirm the effect on the magnitude of the ratio (η), the following three elements were selected.

[0161] The first element is the case where the second spin torque layer is formed of Nb, the second element is the case where the second spin torque layer is formed of Nb and the first spin torque layer is formed of W, and the third element is the case where the second spin torque layer is formed of Nb and the first spin torque layer is formed of Pt.

[0162] In graph 1100, the ratio is (η=-3.03) for the first element, the ratio is (η=-2.18) for the second element, and the ratio is (η=-1.53) for the third element.

[0163] The method for obtaining random numbers from each element is as follows: First, it is confirmed whether SOT switching is ensured under an external magnetic field.

[0164] Then, in an environment without an external magnetic field, a current pulse of sufficient magnitude is injected 1,000 times. If switching occurs, a "1" is input; otherwise, a "0" is input.

[0165] After generating 1,000 random numbers, we verify again whether SOT switching is ensured under an external magnetic field to confirm that the random number sequence was generated from a device that is operating normally.

[0166] The random number sequence generated until each element is no longer able to perform switching operations due to degradation caused by repeated current injection may be 308,683 bits for the first element, 188,406 bits for the second element, and 392,156 bits for the third element.

[0167] The switching probabilities of the elements may be 49.93%, 49.99%, and 49.79%, respectively.

[0168] Graph 1101 shows the results of a minimum entropy test of random number sequences generated by a spin-orbit torque random number generator including a spin torque generation layer including only the second spin torque layer, and a spin-orbit torque random number generator including a spin torque generation layer including the first and second spin torque layers.

[0169] Graph 1101 shows the results of the NIST SP 800-90B test provided by the Institute of Standards and Technology (NIST) to verify the randomness of a random number sequence.

[0170] This test uses 10 different estimators to indicate randomness in terms of entropy, and the lowest value (Min-entropy) among them is displayed as the representative entropy value of the random number sequence, which can be summarized in Table 1 below.

[0171] [Table 1]

[0172] Table 1 shows the test results and minimum entropy of each random number sequence using the Estimator for the first device (Device #1), the second device (Device #2), and the third device (Device #3).

[0173] The first element has a structure in which the spin torque generation layer is formed of Nb, and the second element has a structure in which the spin torque generation layer is formed of a first spin torque layer and a second spin torque layer, the first spin torque layer is formed of W, and the second spin torque layer is formed of Nb.

[0174] Finally, in the third element, the first spin torque layer is made of Pt and the second spin torque layer is made of Nb.

[0175] If the entropy value is n (0≦n≦1), it means that 1 / n operations are required to generate a 1-bit random number, and in the case of a perfect random number generator, the entropy n may be 1.

[0176] The results of the NIST SP800-90B test for each random number sequence can be seen in Figure 10. The Min-entropy values ​​of the three elements are 0.825407, 0.825778, and 0.891471, respectively, ensuring a high entropy value regardless of the magnitude of η.

[0177] Therefore, the present invention can realize a spin-orbit torque random number generator with low switching current while maintaining random number generation function by adjusting the ratio of FLT to DLT and entropy based on the structure of a spin torque layer and a magnetic layer based on a material candidate that increases the ratio of FLT compared to DLT.

[0178] In the above-described specific embodiments, elements included in the invention are expressed as singular or plural by the specific embodiments presented.

[0179] However, the expressions "singular" or "plural" are selected to suit the circumstances presented for the convenience of explanation, and the above-described embodiments are not limited to singular or plural components, and elements expressed as plural may be composed of singular, and elements expressed as singular may be composed of plural.

[0180] Meanwhile, although specific embodiments have been described in the explanation of the invention, it goes without saying that various modifications are possible as long as they do not deviate from the scope of the technical ideas contained in the various embodiments.

[0181] Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined by the following claims as well as equivalents to these claims. [Explanation of symbols]

[0182] 100 Spin-Orbit Torque Random Number Generator 110, 301, 311, 401, 411 First spin torque layer 111, 302, 312, 402, 412 Second spin torque layer 120, 303, 313, 403, 413 Magnetization free layer 121 Magnetization fixed layer 130 Tunnel Barrier Layer 300 1st structure 310 Second structure 304, 314 Damping Torque 305, 315 Field Torque

Claims

1. a spin torque generation layer including a first spin torque layer and a second spin torque layer, a magnetization free layer, a tunnel barrier layer, and a magnetization fixed layer; the second spin torque layer is located between the magnetic free layer and the first spin torque layer and generates a field torque (FLT) and a damping torque (DLT) associated with the magnetization direction of the magnetic free layer; The first spin torque layer adjusts a ratio of a magnitude of the field torque to a magnitude of the damping torque by adjusting a thickness of the second spin torque layer, the magnetization direction of the free magnetic layer is randomly aligned in an up or down direction based on the adjusted ratio of the torques; generating a random number based on a current that is switched and output based on the randomly aligned magnetization directions; When the thickness of the second spin torque layer increases, the torque ratio increases as the magnitude of the damping torque decreases while the magnitude of the field torque is maintained, and when the thickness of the second spin torque layer decreases, the torque ratio decreases as the magnitude of the damping torque increases while the magnitude of the field torque is maintained, the first spin torque layer is formed of a metal having high damping torque efficiency, The second spin torque layer is formed of a metal having a high efficiency of the field torque and a material having a lower efficiency of the damping torque than the first spin torque layer.

2. the first spin torque layer is formed of any one of Ta, W, and Pt; The spin-orbit torque random number generator of claim 1 , wherein the second spin torque layer is made of Nb.

3. The spin-orbit torque random number generator of claim 2, wherein the second spin torque layer is formed to a thickness of 1 nm to 15 nm.

4. 2. The spin-orbit torque random number generator of claim 1, wherein, based on the first spin torque layer and the second spin torque layer, the magnitude of a switching current for switching the magnetization direction is reduced and the switching efficiency of switching the magnetization direction is increased according to a ratio of the field torque component of the spin-orbit torque that is higher than the damping torque component.

5. 5. The spin-orbit torque random number generator of claim 4, wherein the random number generation function is realized by having an entropy value of a random number sequence for a random number based on a current output according to a ratio of the torque to a repeatedly injected current that is higher than a reference value associated with randomness.

6. forming a spin torque generation layer including a first spin torque layer and a second spin torque layer; forming a magnetization free layer on the spin torque generation layer; forming a tunnel barrier layer on the magnetic free layer; forming a magnetization fixed layer on the tunnel barrier layer to form a spin-orbit torque random number generator; the second spin torque layer is located between the magnetic free layer and the first spin torque layer and generates a field torque (FLT) and a damping torque (DLT) associated with the magnetization direction of the magnetic free layer; The first spin torque layer adjusts a ratio of a magnitude of the field torque to a magnitude of the damping torque by adjusting a thickness of the second spin torque layer, the magnetization direction of the free magnetic layer is randomly aligned in an up or down direction based on the adjusted ratio of the torques; the spin-orbit torque random number generator generates a random number based on a current that is switched and output based on the randomly aligned magnetization directions; When the thickness of the second spin torque layer increases, the torque ratio increases as the magnitude of the damping torque decreases while the magnitude of the field torque is maintained, and when the thickness of the second spin torque layer decreases, the torque ratio decreases as the magnitude of the damping torque increases while the magnitude of the field torque is maintained, the first spin torque layer is formed of a metal having high damping torque efficiency, 10. A method for manufacturing a spin-orbit torque random number generator, wherein the second spin torque layer is formed of a metal having a high efficiency of the field torque and a material having a lower efficiency of the damping torque than the first spin torque layer.

7. The step of forming the spin torque generation layer includes: forming the first spin torque layer from one metal selected from Ta, W, and Pt; and forming the second spin torque layer from Nb.

8. 8. The method of claim 7, wherein the second spin torque layer is formed to a thickness of 1 nm to 15 nm.

Images