Multilayer magnetic structure layer resolution test method and system based on dual-mode images

Abstract

The application provides a kind of multi-layer magnetic structure layer resolution test method and system based on dual-mode image, method includes: through the cooperation between reference magnetic state, target magnetic layer and non-target magnetic layer, layer selective write excitation and dual-mode image differential processing, it can attribute the magnetic response change in adjacent magnetic state to corresponding magnetic layer, improve the accuracy of response layer identification in multi-layer magnetic structure;Using the change relationship of MOKE image and NV magnetic field image, distinguish the surface magnetic layer, buried magnetic layer, near compensation magnetic structure response and interlayer coupling response, and obtain the flip sequence, flip interval and magnetic domain evolution process of different magnetic layers, reduce the layer attribution ambiguity caused by single imaging information, improve the resolution ability of different depth, different magnetic response characteristic magnetic layer in multi-layer magnetic structure.

Description

Technical Field

[0001] This invention relates to the field of magnetic characterization technology for multilayer magnetic structures, and in particular to a method and system for layer-resolution testing of multilayer magnetic structures based on dual-modal images. Background Technology

[0002] Spintronic devices such as magnetic random access memory (MRAM), magnetic sensors, magnetic logic devices, and magnetic neuromorphic devices typically employ multilayer magnetic structures. In these structures, different magnetic layers can possess different magnetic anisotropies, coercive fields, saturation fields, and flip regions, and may form complex film stacks through layers of heavy metals, oxides, protective layers, electrodes, or encapsulation layers. Under excitation by external magnetic fields, currents, voltages, light, or microwaves, different magnetic layers may exhibit responses such as magnetic domain nucleation, domain wall motion, magnetization flips, or magnetization rotation, and each layer's response may have different thresholds and evolution sequences.

[0003] Existing magnetic structure testing techniques include global magnetic moment measurement, magneto-optical Kerr effect (MOKE) imaging, and diamond nitrogen-vacancy (NV) color center magnetic imaging. Global magnetic moment measurement can obtain the overall magnetic response of the multilayer magnetic structure under test; MOKE imaging can quickly obtain the magnetic domains and magnetization changes of the surface or near-surface magnetic layers; and diamond NV color center magnetic imaging can obtain the stray magnetic field distribution at the detection plane. These techniques can reflect the magnetic response of a magnetic structure from different perspectives. However, global magnetic moment measurement results include contributions from multiple layers, the response of MOKE imaging is affected by the optical penetration depth and surface signals, and the NV magnetic field image reflects the total stray field of multiple magnetic layers on the detection plane. The stray fields of different magnetic layers may also superimpose, enhance, or cancel each other out. Therefore, in multilayer magnetic structure testing, single test results or parallel images under the same conditions can easily lead to unclear identification of the response source, inaccurate determination of the flipping order, and insufficient identification of interlayer correlation responses.

[0004] Therefore, for spintronic devices or magnetic film stacks containing multiple magnetic layers, it is necessary to establish a layer-resolved testing scheme suitable for multilayer magnetic structures. While maintaining the correspondence between the test area and the magnetic state, the scheme combines magnetic domain information and stray field information obtained by different imaging methods to determine the responding magnetic layer, the flip threshold, the magnetic domain evolution, and the interlayer correlation response.

[0005] This section is intended to provide background or context for the embodiments of this application set forth in the claims. The description herein is not an admission that it is prior art simply because it is included in this section. Summary of the Invention

[0006] One objective of this invention is to provide a method for layer-by-layer resolution testing of multilayer magnetic structures based on dual-modal imaging. By coordinating reference magnetic states, target and non-target magnetic layers, layer-selective writing excitation, and differential processing of dual-modal images, it can attribute changes in magnetic response in adjacent magnetic states to corresponding magnetic layers, improving the accuracy of response layer identification in multilayer magnetic structures. Utilizing the relationship between MOKE images and NV magnetic field images, it distinguishes surface magnetic layers, buried magnetic layers, near-compensated magnetic structure responses, and interlayer coupling responses, and obtains the flipping sequence, flipping intervals, and magnetic domain evolution processes of different magnetic layers, reducing ambiguity in layer attribution caused by single imaging information and improving the resolution capability of magnetic layers with different depths and magnetic response characteristics in multilayer magnetic structures. Another objective of this invention is to provide a system for layer-by-layer resolution testing of multilayer magnetic structures based on dual-modal imaging. A further objective of this invention is to provide a computer-readable medium. A final objective of this invention is to provide a computer device.

[0007] To achieve the above objectives, this invention discloses a method for resolving multilayer magnetic structures based on dual-modal images, comprising: Obtain prior information about the multilayer magnetic structure under test, and determine the target magnetic layer, non-target magnetic layer, and layer selective writing stimuli based on the prior information; A reference magnetic state is constructed based on the target magnetic layer and non-target magnetic layers, and the first dual-mode image of the multilayer magnetic structure under test is acquired under the reference magnetic state. Based on the target magnetic layer and non-target magnetic layers, layer-selective writing excitation is applied, and the second dual-mode image of the multilayer magnetic structure under test is acquired in a stable magnetic state. Layer resolution processing is performed on the first bimodal image and the second bimodal image to generate layer resolution test results.

[0008] Preferably, a reference magnetic state is constructed based on the target magnetic layer and non-target magnetic layers, including: Pre-processing excitation is applied to the multilayer magnetic structure under test to bring the non-target magnetic layer into a stable magnetic state and the target magnetic layer into an initial magnetic state.

[0009] Preferably, the first bimodal image includes a first MOKE image and a first NV magnetic field image; The first dual-mode image of the multilayer magnetic structure under test was acquired under a reference magnetic state, including: The MOKE pattern is determined based on prior information, and wide-field illumination is performed on the multilayer magnetic structure under test based on the MOKE pattern to acquire the first MOKE image. A predetermined measurement bias field is applied to the multilayer magnetic structure under test, and diamond NV centers are excited to collect a preliminary ODMR spectrum. Based on the prepared ODMR spectrum, determine whether the eight ODMR resonance peaks overlap; If so, apply in-plane bias fields along the horizontal and vertical axes, acquire reference ODMR spectra, and calibrate NV axis pairs and crystal orientation relationships based on the reference ODMR spectra; Under a reference magnetic state, an in-plane bias field and a perpendicular bias field are applied to the multilayer magnetic structure under test. The ODMR spectrum of the structure under test is acquired, and a first NV magnetic field image is generated based on the ODMR spectrum of the structure under test, the calibrated NV axis pairs and crystal orientation relationship. If not, perform eight-peak fitting and projection transformation on the prepared ODMR spectrum at each pixel location to obtain the first NV magnetic field image.

[0010] Preferably, layer-selective write stimulation is applied based on the target magnetic layer and non-target magnetic layers, including: A layer-selective writing excitation is applied to the multilayer magnetic structure under test, so that the non-target magnetic layer is within a preset range of variation, causing the target magnetic layer to undergo a change in magnetization state.

[0011] Preferably, layer resolution processing is performed on the first bimodal image and the second bimodal image to generate layer resolution test results, including: Image preprocessing is performed on the first bimodal image and the second bimodal image to obtain a preprocessed bimodal image pair. Image preprocessing includes spatial registration, background subtraction, and intensity normalization. Differential processing is performed on the bimodal image pairs to obtain the image changes between adjacent states; Layer assignment is performed based on the changes in images between adjacent states to obtain the layer resolution test results.

[0012] Preferably, the image changes include MOKE changes and NV changes; Layer assignment is performed based on the image changes between adjacent states to obtain layer resolution test results, including: The layer assignment results are generated by comparing the changes in MOKE and NV. Using the current magnetic state as a reference magnetic state, selectively write the update layer to the stimulus, and continue to execute the step of acquiring the first bimodal image of the multilayer magnetic structure under test in the reference magnetic state; Based on the layer attribution results obtained from multiple updates, layer resolution test results are generated.

[0013] Preferably, layer attribution results are generated by comparing the changes in MOKE and NV, including: If the MOKE change is greater than the preset MOKE change threshold, the NV change is greater than the preset NV change threshold, and the overlap rate between the magnetic domain region corresponding to the MOKE change and the magnetic domain region corresponding to the NV change is greater than the preset overlap rate threshold, the target magnetic layer is determined as the layer assignment result. If the MOKE change is greater than the preset MOKE change threshold, the NV change is greater than the preset NV change threshold, and the overlap rate between the domain region corresponding to the MOKE change and the domain region corresponding to the NV change is less than or equal to the overlap rate threshold, the interlayer coupling response is determined as the layer assignment result. If the MOKE change is greater than the preset MOKE change threshold, and the NV change is less than or equal to the preset NV change threshold, the near-compensated magnetic structure is determined as the layer assignment result. If the MOKE change is less than or equal to the preset MOKE change threshold, and the NV change is greater than the preset NV change threshold, the buried magnetic layer will be determined as the layer assignment result.

[0014] This invention also discloses a multilayer magnetic structure layer resolution testing system based on dual-modal images, applied to the aforementioned multilayer magnetic structure layer resolution testing method. The system includes: The MOKE light source module is used to provide MOKE detection light to the multilayer magnetic structure under test. NV light source module, used to excite the NV color centers of diamond to produce fluorescence; A shared incident and collection optical path module is used to guide the excitation light emitted by the MOKE detection light and / or the NV light source module to the diamond NV sensing structure of the multilayer magnetic structure under test and the in-situ magnetic characterization platform, and to collect the reflected light of the multilayer magnetic structure under test and the fluorescence generated by the diamond NV color center. The microwave excitation module is used to apply a microwave sweep signal to the diamond NV color center to obtain the ODMR spectrum; The in-situ magnetic characterization platform includes a diamond NV sensing structure, a multilayer magnetic structure under test, a magnetic field excitation module, an electrical excitation and readout module, a sample stage, and a displacement module. A diamond NV sensing structure is set on the multilayer magnetic structure under test to detect stray magnetic fields generated by the multilayer magnetic structure under test. The magnetic field excitation module is used to apply preprocessing excitation, layer-selective writing excitation, and measurement bias field to the multilayer magnetic structure under test. The electrical excitation and readout module is used to apply current excitation, voltage excitation, or electrical readout signals to the multilayer magnetic structure under test. The sample stage is used to support the multilayer magnetic structure and / or diamond NV sensing structure to be tested. The displacement module is used to adjust the relative position between the multilayer magnetic structure under test, the microscope objective, and the external magnetic field source, and to make the first bimodal image and the second bimodal image correspond to the same area. The data processing module is connected to the MOKE light source module, the NV light source module, the shared incident and collection optical path module, the microwave excitation module, and the in-situ magnetic characterization platform, respectively. It is used to control the multilayer magnetic structure under test to enter the reference magnetic state; acquire the first bimodal image; apply layer selective writing excitation and acquire the second bimodal image under stable magnetic state; perform layer resolution processing based on the first bimodal image and the second bimodal image to generate layer resolution test results.

[0015] The present invention also discloses a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the method described above.

[0016] The present invention also discloses a computer device, including a memory and a processor, wherein the memory is used to store information including program instructions, and the processor is used to control the execution of the program instructions, wherein the processor executes the program to implement the method described above.

[0017] The present invention also discloses a computer program product, including a computer program / instructions, which, when executed by a processor, implement the method described above.

[0018] This invention acquires prior information about the multilayer magnetic structure under test and determines the target magnetic layer, non-target magnetic layer, and layer-selective writing excitation based on the prior information. Based on the target and non-target magnetic layers, a reference magnetic state is constructed, and a first bimodal image of the multilayer magnetic structure under test is acquired under the reference magnetic state. Based on the target and non-target magnetic layers, a layer-selective writing excitation is applied, and a second bimodal image of the multilayer magnetic structure under test is acquired under a stable magnetic state. Layer-resolution processing is performed on the first and second bimodal images to generate layer-resolution test results. The results are obtained by considering the reference magnetic state and the target magnetic layer. The combination of non-target magnetic layers, layer-selective writing excitation, and dual-modal image differential processing can attribute changes in magnetic response in adjacent magnetic states to the corresponding magnetic layers, improving the accuracy of response layer identification in multilayer magnetic structures. By utilizing the relationship between MOKE images and NV magnetic field images, the surface magnetic layer, buried magnetic layer, near-compensated magnetic structure response, and interlayer coupling response can be distinguished, and the flipping sequence, flipping interval, and magnetic domain evolution process of different magnetic layers can be obtained, reducing the ambiguity of layer attribution caused by single imaging information and improving the resolution of magnetic layers with different depths and magnetic response characteristics in multilayer magnetic structures. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart of a multilayer magnetic structure layer resolution test method based on dual-modal images provided in an embodiment of the present invention; Figure 2 A flowchart of another method for testing the layer resolution of multilayer magnetic structures based on dual-modal images provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of a multilayer magnetic structure layer resolution testing system based on dual-modal images, provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] To facilitate understanding of the technical solution provided in this application, the relevant content of the technical solution will be explained below. A multilayer magnetic structure refers to a film stack, magnetic heterostructure, or magnetic device comprising two or more magnetic layers. Different magnetic layers can have different magnetic anisotropy directions, coercive fields, saturation fields, and flipping ranges. Therefore, under the same or different excitation conditions, each magnetic layer may successively undergo magnetic domain nucleation, domain wall motion, magnetization flipping, or magnetization rotation. If these response differences can be utilized to cause one magnetic layer to undergo a desired change while keeping the other magnetic layers stable, layer-selective testing conditions can be formed.

[0023] MOKE refers to the phenomenon where the polarization state of polarized light changes with the magnetization state of a magnetic material after reflection. MOKE imaging is a magnetic imaging method based on the magneto-optical Kerr effect. When polarized light illuminates a magnetic material and is reflected, the polarization state of the reflected light changes with the magnetization state of the material. By acquiring and analyzing the reflected light, magnetic domain images or magnetization change information corresponding to the selected MOKE mode can be obtained. MOKE imaging features wide field, fast speed, and in-situ readout, and is typically sensitive to surface or near-surface magnetic layers. The magneto-optical Kerr imaging or MOKE imaging in this application includes poloidal, longitudinal, and transverse modes, or combinations thereof.

[0024] Diamond NV centers (NV centers) are point defects in the diamond lattice consisting of substituted nitrogen atoms and adjacent vacancies. Their spin energy levels can be read out optically and microwaveally and respond to external magnetic fields. Diamond NV center magnetic imaging utilizes the sensitivity of the spin energy levels of NV centers in diamond to external magnetic fields for magnetic field imaging. During testing, NV centers are excited by laser and microwave frequency sweep is applied to obtain optically detected magnetic resonance (ODMR) spectra. The magnetic field distribution at the detection plane is then obtained based on the changes in the ODMR resonance frequency. This magnetic field distribution may include one or more magnetic field components of Bx, By, and Bz, or it may include the total field amplitude or other magnetic field characterization quantities.

[0025] The dual-modal images in this application include MOKE images and NV magnetic field images. MOKE images mainly provide information on magnetic domain morphology, domain wall positions, or magnetization changes, while NV magnetic field images mainly provide information on stray magnetic fields generated by the multilayer magnetic structure under test at the diamond NV detection plane. By acquiring dual-modal images in the reference magnetic state and the stable magnetic state after writing, and by performing registration, background subtraction, difference, and change comparison on images of adjacent states, it is possible to determine whether the magnetic response originates from the target magnetic layer, the buried magnetic layer, the near-compensated magnetic structure, or belongs to a multilayer common response or interlayer coupling response.

[0026] The following example, using the data processing module as the execution subject, illustrates the implementation process of the multilayer magnetic structure layer resolution testing method based on dual-modal images provided in this embodiment of the invention. It is understood that the execution subject of the multilayer magnetic structure layer resolution testing method based on dual-modal images provided in this embodiment of the invention includes, but is not limited to, the data processing module.

[0027] Figure 1 A flowchart of a multilayer magnetic structure layer resolution test method based on dual-modal images provided in this embodiment of the invention is shown below. Figure 1 As shown, the method includes: S101. Obtain prior information about the multilayer magnetic structure to be tested, and determine the target magnetic layer, non-target magnetic layer, and layer selective writing stimulus based on the prior information.

[0028] In this embodiment of the invention, the target magnetic layer is the magnetic layer that is expected to undergo magnetic domain nucleation, domain wall movement, magnetization reversal, or magnetization rotation under layer-selective excitation during the current write-hold-read test cycle. The non-target magnetic layer is another magnetic layer whose magnetic state needs to remain unchanged during the current write-hold-read test cycle, or which is only allowed to vary within a predetermined tolerance range. Layer-selective write excitation is a form of excitation set based on the physical response differences of each magnetic layer's coercive field, saturation field, magnetic anisotropy direction, reversal interval, interlayer coupling relationship, and response differences to magnetic field, current, voltage, light, or microwave excitation. Its parameters are configured to cause the target magnetic layer to undergo the expected change while keeping the non-target magnetic layer stable.

[0029] Specifically, prior information about the multilayer magnetic structure under test is acquired. This prior information is used to limit the range of subsequent magnetic state establishment and excitation window selection. Based on the purpose of this test, a target magnetic layer expected to respond is designated from multiple magnetic layers, and one or more non-target magnetic layers that need to remain stable during this write-hold-read cycle are also specified. Based on the differences between the target and non-target magnetic layers in coercive field, magnetic anisotropy direction, or excitation response efficiency, layer-selective write excitation parameters that can drive changes in the target magnetic layer while maintaining the stability of the non-target magnetic layers are determined.

[0030] This step involves identifying the target and non-target magnetic layers before testing, and determining the layer-selective writing excitation parameters that match the target magnetic layer based on prior information about the multilayer magnetic structure under test. This ensures that the subsequent excitation process has a clear target and controlled conditions. This improves the selectivity of the target magnetic layer's response, enhances the stability of the magnetic state of the non-target magnetic layer, provides a definite variable basis for comparing two-modal images of adjacent magnetic states, and improves the repeatability and effectiveness of the layer-resolution testing process.

[0031] S102. Based on the target magnetic layer and the non-target magnetic layer, construct a reference magnetic state, and acquire the first dual-mode image of the multilayer magnetic structure under test in the reference magnetic state.

[0032] In this embodiment of the invention, the first dual-modal image is a combination of images acquired under the same reference magnetic state, including a first MOKE image and a first NV magnetic field image, both corresponding to the same spatial region. The first dual-modal image includes a first MOKE image and a first NV magnetic field image. Specifically, the first MOKE image is a two-dimensional light intensity distribution map reflecting the morphology of magnetic domains in the surface or near-surface magnetic layer, obtained under the reference magnetic state through the magneto-optical Kerr effect; the first NV magnetic field image is a two-dimensional magnetic field map reflecting the total stray field distribution generated by multiple magnetic layers on the detection plane, obtained under the reference magnetic state through diamond NV color center optical detection magnetic resonance.

[0033] In this embodiment of the invention, the reference magnetic state is a known magnetic state that can be repeatedly identified and established by preprocessing the excitation before applying the layer selective writing excitation. This includes, but is not limited to, saturated magnetization state, uniform magnetization state, known magnetic domain state, near-compensated state, or an intermediate state that can be identified by the previous writing step.

[0034] Specifically, based on the identified target and non-target magnetic layers, a preprocessing excitation is applied to establish a reference magnetic state. Under this reference magnetic state, the non-target magnetic layer is in a stable and known magnetic state, while the target magnetic layer is in an initial magnetic state suitable for subsequent selective writing. Under this reference magnetic state, a first MOKE image and a first NV magnetic field image are acquired for the same region of the multilayer magnetic structure under test, or for corresponding regions that can be spatially registered using markers, as benchmarks for subsequent comparisons.

[0035] This step establishes a known reference magnetic state and acquires MOKE and NV magnetic field images under this reference magnetic state, ensuring that subsequent image changes have a unified initial magnetic state and spatial reference. This improves the benchmark consistency of differential comparison between dual-modal images under adjacent magnetic states, enhances the spatial correspondence between magnetic domain changes and stray field changes, and improves the sensitivity of identifying changes in magnetic response during layer-resolution processing.

[0036] S103. Based on the target magnetic layer and non-target magnetic layer, apply layer selective writing excitation and acquire the second dual-mode image of the multilayer magnetic structure under test in a stable magnetic state.

[0037] In this embodiment of the invention, the second bimodal image is a combination of images acquired under stable magnetic conditions, including a second MOKE image and a second NV magnetic field image. Its image type is the same as the first bimodal image, and it is used for pairing and comparison with the first bimodal image. The second bimodal image includes a second MOKE image and a second NV magnetic field image.

[0038] Specifically, based on the target magnetic layer and non-target magnetic layers, a determined layer-selective writing excitation is applied to induce the desired magnetization change in the target magnetic layer. After the writing excitation ends, the excitation source is immediately turned off, and the magnetic state of the multilayer magnetic structure under test is allowed to stabilize. In this stable magnetic state, a second MOKE image and a second NV magnetic field image are acquired again for the spatial region corresponding to S102.

[0039] This step involves bringing the multilayer magnetic structure under test into a stable magnetic state after the layer-selective write excitation is completed, and then acquiring a second bimodal image in this stable magnetic state, thus separating the writing process from the image reading process in time. This improves the consistency of the corresponding magnetic states between the MOKE image and the NV magnetic field image, enhances the comparability of bimodal data within the same write-hold-read cycle, and improves the repeatability and reliability of comparison results between adjacent magnetic states.

[0040] S104. Perform layer resolution processing on the first bimodal image and the second bimodal image to generate layer resolution test results.

[0041] In this embodiment of the invention, layer-resolved processing refers to the process of assigning the observed magnetic response region by region to a specific magnetic layer (surface magnetic layer or buried magnetic layer) by comparing the changes in MOKE images and NV magnetic field images under adjacent magnetic states and combining them with prior information of the multilayer magnetic structure under test. Layer-resolved test results refer to the final output information such as the magnetic layer identifier, the order of flipping of each magnetic layer, the layer-selective excitation threshold or threshold range, the domain nucleation and domain wall evolution characteristics, and information on interlayer synchronous response, correlated response, or stepwise response.

[0042] Specifically, the first and second bimodal images are spatially registered, and the changes in the MOKE image and the NV magnetic field image in adjacent states are compared. Based on the determined writing excitation conditions and prior information about the multilayer magnetic structure under test, and according to the correspondence or difference between the two changes, it is determined whether the response originates from the target magnetic layer, a non-target magnetic layer, or a combined response of multiple layers. This results in a layer-resolved test that includes the attribution of the response layer and the magnetization evolution characteristics.

[0043] This step combines the response information of near-surface magnetic domain changes from MOKE images with the detection information of stray field distribution changes from NV magnetic field images to perform correlation analysis on dual-mode image changes in adjacent magnetic states. This improves the ability to distinguish responses of magnetic layers at different depths, enhances the reliability of identifying multilayer common responses and interlayer correlated responses, and provides layer-resolution basis for the magnetic response characterization and performance analysis of multilayer magnetic devices.

[0044] In the technical solution provided by this invention, prior information of the multilayer magnetic structure under test is obtained, and the target magnetic layer, non-target magnetic layer, and layer-selective writing excitation are determined based on the prior information; a reference magnetic state is constructed based on the target magnetic layer and non-target magnetic layer, and a first bimodal image of the multilayer magnetic structure under test is acquired under the reference magnetic state; a layer-selective writing excitation is applied based on the target magnetic layer and non-target magnetic layer, and a second bimodal image of the multilayer magnetic structure under test is acquired under a stable magnetic state; layer-resolution processing is performed on the first bimodal image and the second bimodal image to generate layer-resolution test results, which are then compared with the reference magnetic state. The combination of target and non-target magnetic layers, layer-selective writing excitation, and dual-modal image differential processing can attribute changes in magnetic response in adjacent magnetic states to the corresponding magnetic layers, improving the accuracy of response layer identification in multilayer magnetic structures. By utilizing the relationship between MOKE images and NV magnetic field images, the surface magnetic layer, buried magnetic layer, near-compensated magnetic structure response, and interlayer coupling response can be distinguished, and the flipping sequence, flipping interval, and magnetic domain evolution process of different magnetic layers can be obtained, reducing the ambiguity of layer attribution caused by single imaging information and improving the resolution of magnetic layers with different depths and magnetic response characteristics in multilayer magnetic structures.

[0045] Figure 2 A flowchart of another method for testing the layer resolution of multilayer magnetic structures based on dual-modal images provided in this invention is shown below. Figure 2 As shown, the method includes: S201. Obtain prior information about the multilayer magnetic structure to be tested, and determine the target magnetic layer, non-target magnetic layer, and layer selective writing stimulus based on the prior information.

[0046] In this embodiment of the invention, each step is executed by the data processing module.

[0047] In this embodiment of the invention, the multilayer magnetic structure is a film stack, magnetic heterostructure, magnetic device, or micro / nano magnetic structure that includes two or more magnetic layers and optional non-magnetic functional layers.

[0048] Prior information refers to information about the multilayer magnetic structure and its magnetic properties obtained before layer-resolved testing of the test multilayer magnetic structure through methods independent of this testing method. This prior information includes, but is not limited to, the number of magnetic layers, the arrangement order of the magnetic layers, the relative depth of each magnetic layer, the type and direction of magnetic anisotropy of each magnetic layer, the value or range of the coercive field of each magnetic layer, the value or range of the saturation field of each magnetic layer, the magnetization reversal range of each magnetic layer, and the interlayer coupling relationships between the magnetic layers. This prior information is derived from the design documents of the multilayer magnetic structure under test, the structural parameters of the film stack, overall magnetic measurement results, conventional MOKE hysteresis loop measurement results, pre-scan test results, or other known characterization results.

[0049] Layer-selective writing excitation is a physical excitation applied to a multilayer magnetic structure. The type and parameters of the layer-selective writing excitation are selected based on the differences in physical response between the magnetic layers, so that under the excitation, the target magnetic layer undergoes a desired change in magnetization state, while the magnetic state changes of non-target magnetic layers are suppressed within their respective allowable ranges. Types of layer-selective writing excitation include at least one of magnetic field excitation, current excitation, voltage excitation, optical excitation, or microwave excitation. Parameters of layer-selective writing excitation include, but are not limited to, at least one of the following: excitation direction, excitation amplitude, excitation duration, number of pulses, pulse interval, or waveform.

[0050] In this embodiment of the invention, prior information is obtained through at least one of the following methods: reading the film stack structure parameters recorded in the design file of the multilayer magnetic structure under test, performing a conventional MOKE hysteresis loop scan on the multilayer magnetic structure under test, conducting a pre-scan experiment on the multilayer magnetic structure under test before formal testing, or indirectly obtaining it from other known characterization results. This prior information acquisition step is used to limit the selection range of the magnetic state type established in subsequent steps and the parameters of the applied layer-selective writing excitation. This prior information is not used to infer the vertical position of each magnetic layer from a single NV magnetic field image.

[0051] Specifically, based on the purpose of this test, a target magnetic layer is selected from the multiple magnetic layers contained in the multilayer magnetic structure. This target magnetic layer is the magnetic layer that is expected to undergo magnetic domain nucleation, domain wall movement, magnetization reversal, or magnetization rotation during the current write-hold-read cycle. From the other magnetic layers in the multilayer magnetic structure besides the target magnetic layer, one or more magnetic layers are selected as non-target magnetic layers. These non-target magnetic layers are those whose magnetic state needs to remain unchanged during the current write-hold-read cycle, or those whose magnetic state is only allowed to change within a predetermined tolerance range. For each non-target magnetic layer, an allowable range of magnetic state variation is set. This allowable range is defined by a numerical range of one or more physical quantities, and monitoring conditions are specified for monitoring whether each non-target magnetic layer remains within the allowable range. The monitoring conditions include at least one of the following monitoring methods: acquiring MOKE images at the discrimination position or region of the non-target magnetic layer's stability for magnetic domain monitoring; acquiring NV magnetic field images at the discrimination position or region for stray field monitoring; and connecting an electrical lead structure at the discrimination position or region for electrical signal monitoring.

[0052] Specifically, based on the differences in coercive field, magnetic anisotropy direction, magnetization reversal interval, and interlayer coupling relationships among the magnetic layers contained in the prior information, the type and parameters of the layer-selective writing excitation are selected. This layer-selective writing excitation is configured such that, after the excitation is applied to the multilayer magnetic structure, the target magnetic layer undergoes the expected magnetic domain nucleation, domain wall movement, magnetization reversal, or magnetization rotation, while the magnetic state changes of each non-target magnetic layer remain within their respective allowable ranges.

[0053] It is worth noting that layer selectivity can be achieved by different coercive fields and flipping intervals, or by different magnetic anisotropy directions, magnetic field angles, pulse numbers, local currents, gate voltages, optical excitation, microwave excitation, temperature changes, or combinations thereof; the excitation form is not limited, but it should be able to change the target magnetic layer and stabilize the non-target magnetic layer.

[0054] This step acquires prior information about the multilayer magnetic structure under test before applying layer-selective writing excitation, and determines the target magnetic layer, non-target magnetic layer, and layer-selective writing excitation parameters based on the same set of prior information. This ensures that the subsequent reference magnetic state construction and writing excitation process have a consistent physical basis. By setting an allowable variation range for the non-target magnetic layer and configuring corresponding stability monitoring conditions, the selectivity of the target magnetic layer response and the determinability of the state retention of the non-target magnetic layer can be improved. This provides stable control variables for comparing bimodal images of adjacent magnetic states, thereby improving the effectiveness of layer attribution determination and the reliability of test data.

[0055] S202. Apply pre-processing excitation to the multilayer magnetic structure to be tested, so that the non-target magnetic layer is in a stable magnetic state and the target magnetic layer is in an initial magnetic state.

[0056] In this embodiment of the invention, the preprocessing stimulus is one or a series of physical stimuli applied to the multilayer magnetic structure under test before applying the layer-selective writing stimulus, in order to establish a reference magnetic state. The form of the preprocessing stimulus is selected according to the specific structure of the multilayer magnetic structure under test and the response characteristics of each magnetic layer, including but not limited to magnetic fields, currents, voltages, light, microwaves, or combinations thereof. The preprocessing stimulus applies to all magnetic layers, but its design goal is to bring non-target magnetic layers into a stable and known magnetic state and to bring the target magnetic layer into an initial magnetic state suitable for subsequent selective writing.

[0057] Specifically, a pre-processing excitation is applied to the multilayer magnetic structure under test. The pre-processing excitation is generated by a corresponding excitation generating device and acts on the multilayer magnetic structure under test. The excitation generating device includes, but is not limited to, at least one of a magnetic field generating coil, an electrical signal generator, a light source, or a microwave signal source.

[0058] The type, direction, amplitude, and duration of the preprocessing excitation are adjusted so that, during or after the application of the preprocessing excitation, the magnetization distribution of the non-target magnetic layer enters a deterministic and repeatable magnetic state that remains stable after the removal of the preprocessing excitation; this is defined as the stable magnetic state of the non-target magnetic layer. Simultaneously, the same preprocessing excitation is adjusted so that the magnetization distribution of the target magnetic layer enters a deterministic initial magnetic state. This initial magnetic state is configured so that, when a layer-selective writing excitation is applied in a subsequent step, the target magnetic layer will undergo a magnetization change in a predetermined direction or pattern. This initial magnetic state can be any of the following: a saturated magnetization state, a uniform magnetization state, a state with a known magnetic domain configuration, a near-compensated state, or a zero-field remanence state, depending on the type of magnetic anisotropy of the target magnetic layer and the current testing objective.

[0059] After the preprocessing excitation achieves its intended effect, it is removed. Following removal, a measurement bias field is retained, applied to the region containing the multilayer magnetic structure under test. Its amplitude is set below a critical threshold that would cause further changes in the magnetic state of any magnetic layer within the multilayer magnetic structure. This measurement bias field is used to maintain the stability of the established magnetic state during subsequent acquisition of the first dual-mode image, or to provide the necessary static magnetic field bias conditions for photodetector magnetic resonance measurements of diamond NV centers. Through the application and removal of the preprocessing excitation, a reference magnetic state is established, in which the non-target magnetic layer is in a stable magnetic state, and the target magnetic layer is in its initial magnetic state.

[0060] It is worth noting that the reference magnetic state can be a saturated state, a uniformly magnetized state, a known magnetic domain state, a near-compensated state, a zero-field remanent state, or an intermediate state obtained from the previous writing step and that can be repeatedly identified. For samples with multiple layers that are sequentially flipped, the state of any completed layer can be used as the reference magnetic state for testing the next magnetic layer.

[0061] This step applies a preprocessing excitation to bring the non-target magnetic layer into a stable and defined magnetic state, and to place the target magnetic layer in a known initial magnetic state, thus providing a clear initial reference for subsequent selective layer writing and comparison of adjacent magnetic states. By retaining a measurement bias field below the magnetization change threshold after removing the preprocessing excitation, the stability of the established magnetic state can be maintained, and the required static magnetic field conditions can be provided for diamond NV center ODMR measurements. This improves the ODMR spectral quality and the repeatability of dual-modal image acquisition without changing the reference magnetic state.

[0062] S203. Determine the MOKE mode based on prior information, and perform wide-field illumination on the multilayer magnetic structure under test based on the MOKE mode to acquire the first MOKE image.

[0063] In this embodiment of the invention, the MOKE mode is a specific working mode of magneto-optical Kerr effect imaging, including poloidal MOKE, longitudinal MOKE, and transverse MOKE. Poloidal MOKE is sensitive to the magnetization component perpendicular to the surface of the multilayer magnetic structure under test, and is suitable for detecting magnetic layers with magnetization directions along the normal to the film surface; longitudinal MOKE is sensitive to the magnetization component parallel to the surface of the multilayer magnetic structure under test and parallel to the light incident plane; transverse MOKE is sensitive to the magnetization component parallel to the surface of the multilayer magnetic structure under test and perpendicular to the light incident plane.

[0064] It is worth noting that, depending on the magnetization direction of the target magnetic layer, poloidal MOKE, longitudinal MOKE, transverse MOKE, or a combination thereof can be used; intensity maps, difference maps, hysteresis loops, time series, or polarization analysis results can be acquired. The selection of the MOKE mode is used to improve the ability to distinguish the stability of target or non-target magnetic layers.

[0065] Specifically, based on the magnetic anisotropy type and magnetization easy axis direction of each magnetic layer contained in the prior information, one or more MOKE imaging modes are selected from a variety of modes for the current acquisition. When the magnetization direction of the magnetic layer related to the current test in the multilayer magnetic structure is perpendicular to the surface of the multilayer magnetic structure under test, the poloidal mode is selected; when the magnetization direction of the magnetic layer is parallel to the surface of the multilayer magnetic structure under test and parallel to the incident plane, the longitudinal mode is selected; when the magnetization direction of the magnetic layer is parallel to the surface of the multilayer magnetic structure under test and perpendicular to the incident plane, the transverse mode is selected.

[0066] It is worth noting that it allows the simultaneous selection of two or more modes to obtain information on different magnetization components.

[0067] The MOKE light source module is activated, generating an incident beam for magneto-optical Kerr imaging. This incident beam is converted into linearly polarized light by a polarizer and expanded by an optical system to form a wide-field illumination spot covering the entire imaging area within the field of view of the multilayer magnetic structure under test. This wide-field illumination spot simultaneously illuminates the surface area of ​​the multilayer magnetic structure under test. The area covered by this wide-field illumination spot is defined as the same spatial area as the subsequent NV magnetic field image acquisition area, or as a corresponding area that can be spatially registered by at least one of the following methods: a registration mark pre-fabricated on the surface of the multilayer magnetic structure under test, the same position recorded by the displacement stage coordinates, or a fixed geometric feature identifiable in the image.

[0068] The light beam reflected from the surface of the multilayer magnetic structure under test carries polarization information from the magneto-optical Kerr effect. After passing through an analyzer, the change in polarization state is converted into a change in light intensity. The image acquisition component exposes and records this information, obtaining a first MOKE image under this reference magnetic state. This first MOKE image serves as one of the benchmarks for spatial registration and variation comparison with the NV magnetic field image corresponding to this state in subsequent steps.

[0069] This step improves the detection sensitivity of MOKE images for target magnetic layers or relevant magnetization components by selecting a matching MOKE mode based on the magnetization direction of each magnetic layer in the prior information, and enhances the quality of magnetic domain images and the reliability of subsequent layer attribution criteria. By acquiring MOKE images using a wide-field illumination method, the magnetic domain distribution within the field of view can be obtained in a single exposure, improving image acquisition efficiency within the stabilization time window after writing. Simultaneously, by ensuring that the MOKE image and the NV magnetic field image correspond to the same region or a registerable region, a unified spatial reference can be provided for subsequent dual-modal image registration, differential analysis, and spatial correspondence determination.

[0070] S204. Apply a predetermined measurement bias field to the multilayer magnetic structure under test, excite the diamond NV color center, and collect the preliminary ODMR spectrum.

[0071] In this embodiment of the invention, the predetermined measurement bias field is a static magnetic field of known direction and amplitude applied to meet the static magnetic field bias conditions required for optically detected magnetic resonance (ODMR) measurements of diamond NV centers. The amplitude of the predetermined measurement bias field is set so as not to change the magnetic state of the multilayer magnetic structure under test, but only to adjust the energy level splitting of the NV centers so that the resonance peak in the ODMR spectrum appears in the observable frequency range.

[0072] Specifically, a predetermined measurement bias field is applied to the multilayer magnetic structure under test. The direction and amplitude of this predetermined measurement bias field are pre-set according to the operating conditions of the diamond NV color center. The NV light source module is activated to excite light and irradiate the NV color center in the diamond NV sensing structure, causing the electron spin state of the NV color center to initialize and undergo the energy level transitions required for readout. Under continuous irradiation by the excitation light, a microwave frequency sweep is performed, i.e., the frequency of the microwave signal applied to the diamond NV color center is gradually changed, and a wide-field NV fluorescence image is acquired at each frequency point. The acquired wide-field NV fluorescence images are processed pixel by pixel, and the curve of fluorescence intensity changing with microwave frequency at each pixel position is extracted to obtain the preliminary ODMR spectrum.

[0073] In this embodiment of the invention, the preliminary ODMR spectrum is a pixel-by-pixel or region-averaged photodetector magnetic resonance spectrum obtained by a single rapid microwave frequency sweep. This spectrum is used to evaluate the separation degree of the eight resonance peaks, that is, to determine whether the eight resonance peaks are separated or overlapped in the frequency domain, so as to select the subsequent data processing path.

[0074] This step, by acquiring preliminary ODMR spectra before full quantitative magnetic field imaging, allows for the prior acquisition of spectral characteristics such as resonance peak distribution, peak spacing, and signal-to-noise ratio under the current measurement conditions. This provides a basis for selecting the eight-peak fitting path and the directional bias field measurement path. This improves the matching between the ODMR data processing path and the actual spectral conditions, enhances the stability and data accuracy of the magnetic field component acquisition process, and improves the automation and reliability of the NV magnetic field image generation workflow.

[0075] S205. Based on the prepared ODMR spectrum, determine whether the eight ODMR resonance peaks overlap. If yes, proceed to S206; otherwise, proceed to S208.

[0076] In this embodiment of the invention, the diamond NV color center has four different crystal orientations. After applying a bias magnetic field, the energy level corresponding to each crystal orientation undergoes Zeeman splitting, with each splitting resulting in two resonance peaks, for a total of eight ODMR resonance peaks. The frequency positions of the eight ODMR resonance peaks are related to the magnitude of the projection of the applied magnetic field onto the corresponding NV axis direction.

[0077] Specifically, the prepared ODMR spectrum is analyzed to extract the frequency positions of all resonance depressions and the full width at half maximum (FWHM) of each resonance depression; the frequency interval between adjacent resonance depressions is calculated, and the minimum value among all adjacent intervals is determined; the minimum frequency interval is compared with the FWHM of each resonance depression; if the minimum frequency interval is less than or equal to a predetermined multiple of the FWHM of the resonance depression, it is determined that the eight ODMR resonance peaks overlap, and S206 is continued; if the minimum frequency interval is greater than the predetermined multiple, it is determined that the eight ODMR resonance peaks do not overlap, that is, they can be sufficiently separated, and S208 is continued.

[0078] It is worth noting that the predetermined multiple is a threshold coefficient set in advance according to the robustness requirements of the fitting algorithm.

[0079] This step determines the corresponding data processing path based on the resonance peak distribution characteristics in the prepared ODMR spectrum, enabling the eight-peak fitting path and the directional bias field measurement path to be applicable to different spectral resolution conditions. This improves the stability of the ODMR spectrum fitting and magnetic field component determination process, allowing NV magnetic field imaging to balance measurement accuracy and efficiency under different signal intensities and peak resolution conditions.

[0080] S206. Apply in-plane bias fields along the horizontal and vertical axes, acquire reference ODMR spectra, and calibrate the NV axis pairs and crystal orientation relationships based on the reference ODMR spectra.

[0081] In this embodiment of the invention, the reference ODMR spectrum includes a first reference wide-field NV fluorescence image and a second reference wide-field NV fluorescence image.

[0082] Specifically, the multilayer magnetic structure under test is brought into a known saturated reference magnetic state, which is obtained by applying a strong magnetic field that saturates all magnetic layers in the same direction. Under this saturated reference magnetic state, a first in-plane bias field of known amplitude is applied along the X-axis of the laboratory coordinate system. A microwave frequency sweep is performed under this bias field, and a first reference wide-field NV fluorescence image is acquired. The in-plane bias field in the X-direction is then removed, and a second in-plane bias field of known amplitude is applied along the Y-axis of the laboratory coordinate system. A microwave frequency sweep is performed again under this bias field, and a second reference wide-field NV fluorescence image is acquired.

[0083] Four-peak fitting was performed on the ODMR spectra of each pixel position in the first and second reference wide-field NV fluorescence images to obtain the frequency splitting variation relationship of the resonance peak before and after applying in-plane bias fields in different directions. Based on this frequency splitting variation relationship, the projection directions of the two sets of NV axis pairs that can be resolved under overlapping conditions were determined, and the angle relationship between the diamond crystal orientation and the X-axis, Y-axis and Z-axis of the laboratory coordinate system was calibrated accordingly, i.e., the crystal orientation relationship.

[0084] In this embodiment of the invention, the saturated reference magnetic state is a known magnetic state achieved by applying a sufficiently strong external magnetic field, causing the magnetization of all magnetic layers in the multilayer magnetic structure to saturate along the same direction. Under this state, the stray field generated by the multilayer magnetic structure under test is deterministic and repeatable, making it suitable as a reference state for calibrating the relationship between the NV crystal orientation and the laboratory coordinate system.

[0085] Four-peak fitting is a linear fitting of the ODMR spectrum to four resonance depressions. This fitting method is suitable for situations where, after applying a bias field of known direction and amplitude, some of the eight peaks are resolved in the frequency domain through pairing, but the overall spectrum only presents four main depressions or four peak clusters. It simplifies the fitting process and improves calibration accuracy.

[0086] Diamond NV color centers have four crystal axes along their body diagonal directions, each constituting an NV orientation. Under an applied bias field, specific bias field directions will cause the resonance peaks corresponding to certain crystal axes to exhibit similar splitting behavior, forming a distinguishable combination, which is called an NV axis pair.

[0087] This step involves applying known bias fields along two orthogonal planes under a known saturated reference state and acquiring corresponding reference ODMR spectra. This allows for the determination of the NV axis pairs and the correspondence between the diamond crystal orientation and the laboratory coordinate system, using a controllable external magnetic field as a calibration benchmark. The calibration results can be used for subsequent ODMR data processing under the desired magnetic state, thereby improving the feasibility and stability of vector magnetic field component determination under eight-peak overlap conditions and enhancing coordinate consistency among multiple rounds of NV magnetic field imaging results.

[0088] S207. Apply in-plane bias field and perpendicular bias field to the multilayer magnetic structure under the reference magnetic state, acquire the ODMR spectrum to be measured, and generate the first NV magnetic field image based on the ODMR spectrum to be measured, the calibrated NV axis pairs and crystal orientation relationship, and continue to execute S209.

[0089] In this embodiment of the invention, the in-plane bias field is a static magnetic field with a known direction and amplitude applied along the X-axis or Y-axis of the laboratory. It is used to split the resonance frequency of the NV color center in a known direction under the condition of eight-peak overlap, so as to separately calculate the in-plane component of the stray field of the multilayer magnetic structure under test from the ODMR spectrum.

[0090] The vertical bias field is a static magnetic field with a known direction and amplitude applied along the Z-axis of the laboratory. It is used to calibrate the response of the NV color center to the out-of-plane magnetic field, thereby solving the out-of-plane component Bz of the stray field of the multilayer magnetic structure under test from the spectral changes.

[0091] Specifically, after calibration, the multilayer magnetic structure under test is placed in a reference magnetic state. When acquiring the first NV magnetic field image, a direction with a relatively high signal-to-noise ratio is selected from the calibrated X-axis and Y-axis directions as the in-plane measurement direction. An in-plane bias field of known amplitude is applied along this direction, and the ODMR spectrum under test is acquired under this condition. Based on the calibrated NV axis pairs and the angle relationship between the diamond crystal orientation and the laboratory coordinate system (crystal orientation relationship), the contribution of the applied known in-plane bias field and the system background field is subtracted from the resonance frequency shift of the ODMR spectrum under test to obtain the two-dimensional distribution of the in-plane magnetic field components Bx and By generated by the multilayer magnetic structure under test.

[0092] For the out-of-plane magnetic field component Bz, a vertical bias field of known amplitude is applied along the Z-axis of the laboratory coordinate system, and the ODMR spectrum to be measured is acquired under the reference magnetic state. The ODMR spectrum to be measured is compared pixel by pixel with the reference ODMR spectrum acquired under the same vertical bias field in the saturated reference magnetic state, and the change in resonance peak frequency or frequency splitting under the two states is extracted. The contributions of the vertical bias field and the system background field are subtracted from the change to obtain the two-dimensional distribution of the out-of-plane magnetic field component Bz generated by the multilayer magnetic structure to be measured. The three components Bx, By and Bz are combined into a three-dimensional vector field as the first NV magnetic field image.

[0093] In this embodiment of the invention, the first NV magnetic field image corresponds to the total stray field distribution generated by each magnetic layer under the reference magnetic state, reflecting the total stray field vector distribution generated by all magnetic layers in the multilayer magnetic structure at the diamond NV color center detection plane.

[0094] It is worth noting that when the eight ODMR peaks can be distinguished, eight-peak fitting is used; when the eight peaks overlap, directional bias field measurement is used; other measurement sequences that can calibrate the NV axis projection and complete coordinate transformation can also be used according to the sample and diamond crystal orientation, but the bias field and background field should be deducted in all cases.

[0095] This step employs a directional bias field strategy, separating the NV axial relationship calibration process from the magnetic field component measurement process under the tested magnetic state, and separately calculating the in-plane and out-of-plane magnetic field components based on the calibrated coordinate correspondence. This improves the stability and completeness of the Bx, By, and Bz vector field acquisition under eight-peak overlap conditions, providing reliable vector magnetic field constraints for subsequent quantitative tomographic resolution and magnetization reconstruction. Simultaneously, the calibration results can be reused under multiple tested magnetic states, thereby improving the testing efficiency of multi-round NV magnetic field imaging.

[0096] S208. Perform eight-peak fitting and projection transformation on the prepared ODMR spectrum at each pixel location to obtain the first NV magnetic field image.

[0097] In this embodiment of the invention, eight-peak fitting is a data processing method for performing line fitting on eight independently resolvable resonance depressions in an ODMR spectrum. The line shape function used in the fitting is either a Lorentz function or a Gaussian function, and the fitting output is the precise center frequency of each resonance peak. This method is suitable for spectral conditions where the eight resonance peaks are sufficiently separated in the frequency domain and the frequency spacing between adjacent peaks is significantly greater than the full width at half maximum (FWHM) of each peak.

[0098] Projection transformation is a mathematical operation that transforms the projected values ​​of the magnetic field vector from the four body diagonal crystal axes of the NV color center into components along the three orthogonal directions of the X, Y, and Z axes in a laboratory Cartesian coordinate system through a coordinate rotation matrix. This transformation relies on the fixed crystal orientation angle relationships in the diamond crystal structure.

[0099] Specifically, for a stack of wide-field NV fluorescence images acquired under a reference magnetic state, a preliminary ODMR spectrum is extracted for each pixel location it covers. This spectrum is then fitted with eight peaks to precisely locate the frequency centers of the eight resonance peaks. The eight resonance peaks are divided into four groups according to their frequency splitting and pairing relationships, with each group corresponding to a crystal orientation of the diamond NV color center. This yields the projection values ​​of the magnetic field along the four NV axes.

[0100] Based on the known directional relationships between the four NV crystal directions in the diamond crystal structure and the laboratory coordinate system, a projection transformation is performed to convert the four NV axis projection values ​​at each pixel position into X-axis, Y-axis, and Z-axis components in the laboratory coordinate system. The contributions of the applied measurement bias field and the system background field are subtracted from the three transformed components to obtain the two-dimensional distribution of Bx, By, and Bz generated by the multilayer magnetic structure under test, which serves as the first NV magnetic field image.

[0101] Under the condition that all eight ODMR resonance peaks are resolvable, this step obtains the magnetic field projections corresponding to the four NV orientations through a single wide-field microwave frequency sweep and pixel-by-pixel eight-peak fitting, and then obtains the three-component vector magnetic field distributions of Bx, By, and Bz through coordinate transformation. This can improve the acquisition efficiency and testing throughput of vector NV magnetic field images in scenarios with high spectral quality, and enhance the self-consistency between the three-component magnetic field results.

[0102] S209. Apply layer-selective writing excitation to the multilayer magnetic structure under test, so that the non-target magnetic layer is within a preset range of variation, causing the target magnetic layer to undergo a change in magnetization state.

[0103] Specifically, based on the type and parameters of the layer-selective writing excitation, a corresponding excitation application device is used to apply the layer-selective writing excitation to the multilayer magnetic structure under test. During the application of the layer-selective writing excitation, the magnetic state changes of the non-target magnetic layers are monitored in real time or determined afterward, according to the set allowable change ranges for each non-target magnetic layer, to ensure that the changes are suppressed within their respective allowable change ranges. The layer-selective writing excitation is configured such that, during or after the excitation, the target magnetic layer undergoes at least one of the following magnetization state changes: domain nucleation, domain wall movement, magnetization reversal, or magnetization rotation, while the amount of magnetic state change of each non-target magnetic layer does not exceed the corresponding set allowable change range.

[0104] In this embodiment of the invention, the type of magnetization state change depends on the magnetic anisotropy type of the target magnetic layer, the current initial magnetic state, and the direction and amplitude of the applied excitation. After applying layer-selective write excitation, if it is determined that the change of a non-target magnetic layer exceeds its allowable range, the test result of the current write-hold-read cycle is marked as single-layer selectivity not being achievable under the excitation parameters, and this result is recorded for subsequent analysis.

[0105] As an alternative, the excitation application device is selected according to the excitation type: when the excitation type is a magnetic field, a magnetic field generating coil or electromagnet is used; when the excitation type is current, an electrical signal generator is used and current is injected through an electrode structure connected to the multilayer magnetic structure under test; when the excitation type is voltage, a voltage source is used and voltage is applied through an electrode structure connected to the multilayer magnetic structure under test; when the excitation type is light, a pulsed laser is used and the light spot is adjusted to cover the target area; when the excitation type is microwave, a microwave signal source is used and radiated to the multilayer magnetic structure under test through a microwave transmission structure.

[0106] In this embodiment of the invention, the change in magnetization state is the behavior of the target magnetic layer under the excitation of layer selective writing, including domain nucleation (forming new reverse domains in the originally uniformly magnetized region), domain wall movement (the existing domain boundary is displaced, causing the domain size to change), magnetization reversal (the magnetization direction is switched from one direction to the opposite direction), and magnetization rotation (the magnetization vector direction is deflected at an angle in the plane or in space).

[0107] It is worth noting that the specific manifestation of the response type to the change in magnetization state depends on the type of magnetic anisotropy of the target magnetic layer and the form of the applied excitation.

[0108] This step applies excitation to the multilayer magnetic structure under test according to predetermined layer-selective writing excitation parameters. This allows for controlled driving of the target magnetic layer to produce the expected magnetization response, while keeping non-target magnetic layers within a preset range of variation. This provides a clear excitation variable for subsequent comparisons of adjacent magnetic states, improving the controllability of the layer-resolved testing process and the validity of the layer attribution results.

[0109] S210. Acquire the second dual-mode image of the multilayer magnetic structure under test in a stable magnetic state.

[0110] In this embodiment of the invention, the second bimodal image includes a second MOKE image and a second NV magnetic field image.

[0111] In this embodiment of the invention, a stable magnetic state refers to an equilibrium state in which the magnetization distribution of the multilayer magnetic structure under test no longer changes significantly over time after a sufficiently long relaxation time following the shutdown of the write excitation. After reaching a stable magnetic state, the domain configuration and magnetization direction of the multilayer magnetic structure under test remain essentially unchanged during subsequent image acquisition, thereby ensuring that MOKE images and NV magnetic field images acquired at different times correspond to the same magnetic state.

[0112] Specifically, after the layer-selective writing excitation is applied, the excitation source of the writing excitation is immediately turned off, so that the multilayer magnetic structure under test is no longer subject to active excitation that continues to drive changes in magnetization. After turning off the excitation source, a predetermined time is waited for the magnetic state of the multilayer magnetic structure under test to stabilize. This predetermined time is preset according to the magnetic relaxation characteristics of the multilayer magnetic structure under test, or determined by real-time monitoring of the magnetic state signal until the signal no longer changes significantly over time.

[0113] After confirming that the magnetic state has reached stability, and without applying any further excitation to drive the magnetization change of the multilayer magnetic structure under test, a second MOKE image and a second NV magnetic field image are acquired for the same spatial region as when the first dual-mode image was acquired.

[0114] It is worth noting that when acquiring the second MOKE image, the operations in S203—selecting the MOKE mode based on prior information, performing wide-field illumination, and recording the reflected light image—are repeated, and the selection of the MOKE mode remains the same as in S203. Further details of this invention will not be elaborated upon here.

[0115] It is worth noting that when acquiring the second NV magnetic field image, under the NV measurement conditions corresponding to the reference magnetic state, a bias field for optically probed magnetic resonance measurement is applied. The amplitude of this bias field is set below the critical threshold that causes the magnetization state of the multilayer magnetic structure under test to continue to change or evolve further. Under this bias field condition, the ODMR spectral acquisition and magnetic field component conversion operations corresponding to paths S204 to 208 are repeated to obtain the second NV magnetic field image. Further details of this invention will not be elaborated upon here.

[0116] It is worth noting that the acquisition of MOKE images and NV magnetic field images should be performed according to any of the following timing methods: sequential acquisition, synchronous acquisition, or quasi-synchronous acquisition, but all must correspond to the same stable magnetic state after the write excitation is turned off. Quasi-synchronous acquisition means that the MOKE image acquisition and NV magnetic field image acquisition are not performed at the same time, but the time interval between the two acquisitions is shorter than the time required for a detectable change in the magnetic state of the multilayer magnetic structure under test, thus allowing the two sets of images to still be considered to correspond to the same magnetic state.

[0117] This step involves acquiring a second bimodal image after the layer-selective writing excitation is turned off and a stable magnetic state is formed. This ensures that the MOKE image, NV magnetic field image, and data at each frequency point during the NV sweep correspond to the same post-writing magnetic state. By controlling the NV measurement bias field below the critical threshold of magnetization change, the stability of the measured magnetic state during image acquisition can be maintained, improving the consistency of states between bimodal images and within the NV sweep data, and enhancing the accuracy of differential changes between adjacent magnetic states.

[0118] S211. Perform image preprocessing on the first bimodal image and the second bimodal image to obtain a preprocessed bimodal image pair.

[0119] In this embodiment of the invention, image preprocessing includes spatial registration, background subtraction, and intensity normalization, which can improve comparison accuracy and eliminate non-magnetic interference.

[0120] Specifically, spatial registration is performed on the first bimodal image and the second bimodal image, which involves aligning the coordinates of the first MOKE image and the second MOKE image, the first NV magnetic field image and the second NV magnetic field image, the first MOKE image and the first NV magnetic field image, and the second MOKE image and the second NV magnetic field image at the same spatial position, respectively, at the pixel level or at the sub-pixel level.

[0121] As an alternative, spatial registration is achieved through at least one of the following methods: coordinate matching using registration marks pre-fabricated on the surface of the multilayer magnetic structure under test, offset correction using displacement coordinate records of the sample stage, or image cross-correlation registration by extracting identifiable fixed geometric features from the image.

[0122] After spatial registration is completed, background subtraction is performed on the images: the corresponding non-magnetic background components are subtracted from each MOKE image and each NV magnetic field image to eliminate non-magnetic signal components such as optical system inhomogeneity, illumination spot inhomogeneity and dark noise of the detection system.

[0123] As an alternative, the background component can be obtained by: acquiring a reference image at a location where the magnetic state of the multilayer magnetic structure under test does not contribute to the signal, acquiring the system background under conditions where there is no multilayer magnetic structure under test, or extracting a region from the image that does not contain magnetic signals as a background reference.

[0124] The intensity of the image after background removal is normalized. Specifically, the pixel intensity values ​​of each image are scaled according to a unified reference standard to eliminate cross-image intensity differences caused by fluctuations in light source intensity, differences in camera gain, or slight changes in exposure time, so that the pixel intensity between different images and between different magnetic states is comparable.

[0125] After spatial registration, background subtraction, and intensity normalization, the first bimodal image and the second bimodal image form a preprocessed bimodal image pair.

[0126] This step spatially registers the first and second bimodal images to maintain the spatial correspondence between the MOKE and NV magnetic field images under adjacent magnetic states, thereby improving the spatial accuracy of the differential comparison. By performing background subtraction and intensity normalization on the bimodal images, the influence of optical background, system drift, and changes in acquisition conditions on the calculation of image changes can be reduced, improving the accuracy and reliability of MOKE and NV change measurements.

[0127] S212. Perform differential processing on the dual-modal image pair to obtain the image change amount between adjacent states.

[0128] In this embodiment of the invention, the image change includes MOKE change and NV change.

[0129] In this embodiment of the invention, the pixel-by-pixel difference between the second MOKE image and the first MOKE image is calculated to obtain the MOKE change; the pixel-by-pixel difference between the second NV magnetic field image and the first NV magnetic field image is calculated to obtain the NV change. Both the MOKE change and the NV change are two-dimensional spatial distribution maps, and the value of each pixel position in the map represents the amplitude and direction of the signal change between two adjacent magnetic states at that position.

[0130] By subtracting pixel by pixel, the intensity or magnetic field values ​​of pixels at the same coordinate position are subtracted one by one to obtain the MOKE change and NV change.

[0131] In this embodiment of the invention, S212 specifically includes: S2121. Based on the comparison of MOKE change and NV change, generate layer assignment results.

[0132] In this embodiment of the invention, a MOKE change threshold and an NV change threshold are set. These two thresholds are predetermined based on the noise levels and background fluctuations of the first bimodal image and the second bimodal image, respectively. Pixel regions whose MOKE change amplitude exceeds the preset MOKE change threshold are marked as MOKE response regions; pixel regions whose NV change amplitude exceeds the preset NV change threshold are marked as NV response regions.

[0133] Calculate the domain overlap rate between the MOKE response region and the NV response region. The domain overlap rate is the ratio obtained by dividing the intersection area of ​​the MOKE response region and the NV response region by the union area of ​​the two regions.

[0134] Based on the comparison results of MOKE change and MOKE change threshold, the comparison results of NV change and NV change threshold, and the magnetic domain overlap rate between the MOKE response region and the NV response region, the layer assignment results are generated according to the following rules: First scenario: If the MOKE change is greater than the preset MOKE change threshold, the NV change is greater than the preset NV change threshold, and the overlap rate between the magnetic domain region corresponding to the MOKE change and the magnetic domain region corresponding to the NV change is greater than the preset overlap rate threshold, the target magnetic layer in the current write-hold-read cycle is determined as the layer assignment result.

[0135] The second scenario: If the MOKE change is greater than a preset MOKE change threshold, the NV change is greater than a preset NV change threshold, and the overlap rate between the magnetic domain regions corresponding to the MOKE change and the magnetic domain regions corresponding to the NV change is less than or equal to the overlap rate threshold, the interlayer coupling response is determined as the layer attribution result. The interlayer coupling response includes the presence of multiple common responses or interlayer correlated responses, indicating that although changes were detected in both imaging channels, the spatial distribution of the changes is inconsistent, suggesting the presence of changes in multiple magnetic layers or the existence of non-negligible interlayer interactions.

[0136] Third scenario: If the MOKE change is greater than a preset MOKE change threshold, and the NV change is less than or equal to a preset NV change threshold, the near-compensated magnetic structure is determined as the layer assignment result. This scenario is common when the target magnetic layer is a synthetic antiferromagnetic structure (SAF) composed of two sublayers with opposite magnetic moment directions. In the near-compensated state, its magnetization change can be detected in the MOKE image, but the far-field stray field it generates is significantly weakened at the NV detection plane due to mutual cancellation. SAF refers to a structure composed of at least two magnetic sublayers with opposite magnetic moment directions and antiferromagnetic correlation arrangement through a spacer layer; Embodiment 3 of this application is used to distinguish the top SAF as a whole from the buried magnetic layer below it, and is not used to distinguish the two sublayers inside the SAF.

[0137] Fourth scenario: If the MOKE change is less than or equal to a preset MOKE change threshold, and the NV change is greater than a preset NV change threshold, the buried magnetic layer is determined as the layer assignment result. A buried magnetic layer is a magnetic layer located beneath other films, capping layers, electrodes, encapsulation layers, or surface magnetic layers, causing its MOKE response to attenuate, being masked, or invisible. This scenario indicates that the change occurs at a location where MOKE imaging has insufficient optical sensitivity to the surface or near-surface, suggesting that the response originates from a deeper layer within the tested multilayer magnetic structure or a magnetic layer obscured by the upper structure.

[0138] It is worth noting that the judgment of the above four scenarios is based on the premise that the non-target magnetic layer remains within its respective allowable range of variation during the layer-selective write excitation process. If this premise is not met, the test result of the current write-hold-read cycle will be judged as write-selective failure and will not be included in the above layer classification.

[0139] This step transforms the changes in MOKE, NV, and their spatial correspondence into quantifiable layer assignment criteria by setting MOKE variation thresholds, NV variation thresholds, and overlap rate thresholds. This improves the objectivity, automation, and repeatability of layer-resolution test results. Adjusting the thresholds according to the type of the multilayer magnetic structure under test and the measurement conditions enhances the adaptability of the decision rules to different multilayer magnetic structures. Furthermore, distinguishing between spatially corresponding and spatially non-corresponding changes using the overlap rate threshold improves the accuracy of distinguishing between the target magnetic layer response and interlayer correlation responses, providing a more refined basis for analyzing interlayer interactions in multilayer magnetic structures.

[0140] S2122. Using the current magnetic state as the reference magnetic state, selectively write the update layer to the stimulus, and continue to execute the step of acquiring the first dual-mode image of the multilayer magnetic structure under test in the reference magnetic state.

[0141] In this embodiment of the invention, the stable magnetic state at the end of the current write-hold-read cycle is set as the new reference magnetic state for the next cycle. When setting the new reference magnetic state, the preprocessing stimulus is not reapplied; instead, the magnetic state corresponding to the current second bimodal image is directly used as the reference magnetic state for the next cycle, thereby achieving a rolling update of the reference magnetic state.

[0142] The parameters for selective layer writing excitation are updated based on the layer assignment results, including at least one of the following parameters: excitation direction is reversed, excitation amplitude is increased or decreased by a predetermined step size, duration of a single pulse is changed, number of pulses in a pulse sequence is increased or decreased, time interval between adjacent pulses is adjusted, or excitation waveform is switched from one form to another. The direction and step size of parameter adjustment are determined based on the following objectives: to continue obtaining the gradual magnetization evolution of the same target magnetic layer under different excitation intensities to determine its reversal threshold range, or to switch the target magnetic layer to another magnetic layer that has not yet been tested to obtain its response sequence.

[0143] After completing the update of the reference magnetic state and the excitation parameters, return to execute S203, that is, based on the updated reference magnetic state, acquire the first bimodal image under the new reference magnetic state, and continue to execute the subsequent steps to start the next write-hold-read cycle.

[0144] This step uses the current magnetic state as the reference magnetic state for the next round of testing, and continues the dual-modal image acquisition and layer-selective writing process based on the updated reference magnetic state. This reduces the operation time required for repeated resets and re-establishing the reference state, improving the efficiency of multi-round layer-resolution testing. By continuously updating the reference magnetic state, a continuous evolution sequence between adjacent magnetic states can be formed, improving the sampling integrity of magnetic layer flip intervals, magnetic domain evolution processes, and interlayer response timing.

[0145] S2123. Based on the layer attribution results obtained from multiple updates, generate layer resolution test results.

[0146] In this embodiment of the invention, after obtaining the layer assignment results for multiple write-hold-read cycles, the layer assignment results for each cycle are comprehensively organized. The layer selective write stimulus parameters corresponding to each cycle are associated with the corresponding layer assignment results, and the response sequences of each cycle are arranged according to the order of change of the stimulus parameters.

[0147] In this embodiment of the invention, the layer resolution test results include, but are not limited to, the layer affiliation identifier, the order of flipping, the excitation threshold or threshold range, the magnetic domain evolution process, the magnetization rotation path, and the interlayer response relationship.

[0148] Based on the above response sequence, at least one of the following layer-resolved test results is generated: the order of flipping of each magnetic layer (arranged according to the order in which each layer first shows a response during the change of excitation parameters); the layer-selective excitation threshold or threshold range of each magnetic layer (the minimum excitation parameter value required for each layer to begin to produce a detectable response, and the parameter range corresponding to the start of the response to response saturation); the domain nucleation location and domain wall evolution path of each magnetic layer (the spatial evolution process of the magnetic domain morphology as the excitation parameters changes, extracted from multiple cycles of MOKE changes); the magnetization rotation trajectory of each magnetic layer (the magnetization direction change path analyzed from multiple cycles of NV change vector field); and the synchronous response, correlated response, or stepwise response exhibited by each magnetic layer in the excitation sequence (determined based on the temporal relationship of multiple layer assignment results in the dimension of excitation parameters).

[0149] When the NV magnetic field image acquisition process provides Bx, By, and Bz three-component vector data, the quantitative analysis section further outputs the magnetization intensity distribution, magnetization direction spatial distribution, and independent contribution of each magnetic layer to the total stray field of the detection plane, obtained through vector constraints. Qualitative layer resolution is a test to determine whether the response originates from a surface magnetic layer, a buried magnetic layer, or a joint response of multiple magnetic layers, based on the excitation conditions of the target and non-target magnetic layers and the relationship between the MOKE image and the NV magnetic field image; this test does not necessarily require simultaneous acquisition of Bx, By, and Bz. Quantitative layer resolution, based on the qualitative layer attribution, utilizes vector magnetic field constraints such as Bx, By, and Bz, along with necessary magnetic field propagation or magnetization reconstruction, to determine the magnetization distribution, magnetization intensity, or stray field contribution of each magnetic layer.

[0150] It is worth noting that NV data can be a complete ODMR magnetic field image, or it can be a fixed-frequency fluorescence change, a selected magnetic field component, the total field amplitude, or other calibrated magnetic field characterization quantities. Qualitative tomography is not limited to using all three components.

[0151] Furthermore, the above layer resolution test results are output as structured data or in report form.

[0152] This step involves serializing the layer assignment results obtained from multiple cyclic tests and combining them with the layer-selective writing excitation parameters corresponding to each round of testing. This establishes a correspondence between the excitation parameters and the magnetic layer responses. This improves the completeness of the characterization of the response thresholds, response sequences, and magnetic domain evolution processes of each magnetic layer, and provides data support for the layer-by-layer control performance analysis and interlayer interaction analysis of multilayer magnetic structures.

[0153] This invention improves the accuracy and repeatability of determining the response magnetic layer in multilayer magnetic structures by establishing a reference magnetic state, identifying target and non-target magnetic layers, applying layer-selective writing excitation, and comparing bimodal image changes under adjacent magnetic states. This establishes the attribution of the response layer based on the combined constraints of controlled excitation conditions and image change relationships. By jointly utilizing the magnetic domain morphology information provided by MOKE images and the stray field change information provided by diamond NV center magnetic field images, the invention enhances the ability to distinguish between surface magnetic layers, buried magnetic layers, near-compensated magnetic structures, and interlayer correlation responses. It is applicable to various multilayer magnetic structures, including perpendicular magnetic anisotropy / perpendicular magnetic anisotropy (PMA / PMA), perpendicular magnetic anisotropy / in-plane magnetic anisotropy (PMA / IMA), SAF, and buried magnetic layers. By separating the writing excitation from the image reading process and making the MOKE image and NV magnetic field image under adjacent magnetic states correspond to the same region or a registerable region, the state consistency and spatial consistency of the differential comparison of dual-modal images can be improved. When quantitative analysis is required, the Bx, By, and Bz vector magnetic field components can also be obtained through eight-peak fitting or directional bias field measurement, providing reliable vector constraints for magnetization reconstruction and multi-magnetic source separation.

[0154] It is worth noting that the acquisition, storage, use, and processing of data in the technical solution of this application all comply with relevant laws and regulations. The user information in the embodiments of this application was obtained through legal and compliant means, and the acquisition, storage, use, and processing of user information have been authorized and agreed upon by the client.

[0155] It is worth noting that the information collected in this application is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant countries and regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation portals are provided for users to choose to authorize or refuse.

[0156] It is worth noting that the technical solution provided in this application provides users with a corresponding operation entry point, allowing users to choose to agree to or reject the automated decision-making result; if the user chooses to reject, the process will proceed to the expert decision-making process.

[0157] The technical solution of the layer-resolution testing method for multilayer magnetic structures based on dual-modal images provided in this invention involves: acquiring prior information about the multilayer magnetic structure under test; determining the target magnetic layer, non-target magnetic layer, and layer-selective writing excitation based on the prior information; constructing a reference magnetic state based on the target and non-target magnetic layers; acquiring a first dual-modal image of the multilayer magnetic structure under test under the reference magnetic state; applying a layer-selective writing excitation based on the target and non-target magnetic layers; and acquiring a second dual-modal image of the multilayer magnetic structure under test under a stable magnetic state; and performing layer-resolution processing based on the first and second dual-modal images to generate a layer-resolution test result. The results show that by coordinating reference magnetic states, target and non-target magnetic layers, layer-selective writing excitation, and dual-modal image differential processing, changes in magnetic response in adjacent magnetic states can be attributed to the corresponding magnetic layers, improving the accuracy of response layer identification in multilayer magnetic structures. By utilizing the relationship between MOKE images and NV magnetic field images, the surface magnetic layer, buried magnetic layer, near-compensated magnetic structure response, and interlayer coupling response can be distinguished. The flipping sequence, flipping interval, and magnetic domain evolution process of different magnetic layers can be obtained, reducing ambiguity in layer attribution caused by single imaging information and improving the resolution of magnetic layers with different depths and magnetic response characteristics in multilayer magnetic structures.

[0158] Figure 3 This is a schematic diagram of a multilayer magnetic structure layer resolution testing system based on dual-modal images, provided in an embodiment of the present invention. This system is applied to the aforementioned multilayer magnetic structure layer resolution testing method based on dual-modal images, such as... Figure 3 As shown, the system includes: MOKE light source module 100, NV light source module 200, common incident and collection optical path module 300, microwave excitation module 400, in-situ magnetic characterization platform 500 and data processing module 600.

[0159] The MOKE light source module 100 is used to provide MOKE detection light to the multilayer magnetic structure 502 under test.

[0160] The NV light source module 200 is used to excite the NV color centers of diamond to produce fluorescence.

[0161] The shared incident and collection optical path module 300 is used to guide the excitation light emitted by the MOKE detection light and / or the NV light source module to the diamond NV sensing structure of the multilayer magnetic structure under test and the in-situ magnetic characterization platform, and to collect the reflected light of the multilayer magnetic structure under test and the fluorescence generated by the diamond NV color center.

[0162] The microwave excitation module 400 is used to apply a microwave sweep signal to the diamond NV color center to obtain the ODMR spectrum.

[0163] The in-situ magnetic characterization platform 500 includes a diamond NV sensing structure 501, a multilayer magnetic structure under test 502, a magnetic field excitation module 503, an electrical excitation and readout module 504, a sample stage 505, and a displacement module 506.

[0164] The data processing module 600 is connected to the MOKE light source module 100, the NV light source module 200, the common incident and collection optical path module 300, the microwave excitation module 400, and the in-situ magnetic characterization platform 500, respectively. It is used to control the multilayer magnetic structure under test to enter the reference magnetic state; acquire the first bimodal image; apply layer selective writing excitation and acquire the second bimodal image under stable magnetic state; and perform layer resolution processing based on the first bimodal image and the second bimodal image to generate layer resolution test results.

[0165] like Figure 3 As shown, the MOKE light source module 100 includes a MOKE light source 101, a first lens 102, a first aperture 103, a second lens 104, and a second aperture 105.

[0166] The light emitted from the MOKE light source 101 passes sequentially through the first lens 102, the first aperture 103, the second lens 104, and the second aperture 105 before entering the common incident and collection optical path module 300. The first lens 102 and the second lens 104 are used to converge, collimate, expand, or image the MOKE detection light; the first aperture 103 and the second aperture 105 are used to adjust the beam size, incident range, and illumination uniformity.

[0167] As an optional solution, the MOKE light source 101 can be an LED light source, a laser light source, a white light source, or a monochromatic light source.

[0168] like Figure 3 As shown, the NV light source module 200 is used to excite the NV color centers in the diamond NV sensing structure 501 to generate fluorescence. The NV light source module 200 includes a laser 201, a beam homogenization element 202, and a third lens 203.

[0169] As an alternative, laser 201 can be a 532nm green laser or other lasers capable of exciting NV color center fluorescence. Beam homogenization element 202 can be a laser speckle attenuator, a rotating diffuser, a vibrating fiber, a multimode fiber, or an integrating rod, used to improve the illumination uniformity of the excitation light. Third lens 203 is used to collimate, expand, or focus the excitation light.

[0170] It is worth noting that MOKE imaging and diamond NV color center magnetic imaging can be performed by using the same objective lens and the same camera to acquire data sequentially, or by using different objectives lens, different cameras, or independent optical paths to acquire data of the same area or corresponding areas with stable registration marks; as long as adjacent magnetic states can be reliably registered, this method can be performed.

[0171] like Figure 3 As shown, the common incident and collection optical path module 300 is used to guide the MOKE detection light and / or NV excitation light to the in-situ magnetic characterization platform 500, and to collect the reflected light of the multilayer magnetic structure 502 under test and the fluorescence generated by the diamond NV sensing structure 501.

[0172] The shared incident and collection optical path module 300 includes a switchable reflector 301, a fourth lens 302, a first polarization control element 303, a microscope objective 304, a beam splitter 305, a second polarization control element 306, a fifth lens 307, a dichroic mirror 308, a filter assembly 309, and an image acquisition assembly 310.

[0173] The switchable mirror 301 is used to select the MOKE detection light, NV excitation light, or both to enter a common optical path.

[0174] The fourth lens 302 is used to collimate, expand, or focus the light beam entering the common optical path.

[0175] The first polarization control element 303 is used to adjust the polarization state of the MOKE detection light incident on the multilayer magnetic structure 502 to be tested.

[0176] The microscope objective 304 is used to focus or project the light beam onto the area where the multilayer magnetic structure 502 and the diamond NV sensing structure 501 are located, and to collect the reflected light of the multilayer magnetic structure 502 and the fluorescence of the diamond NV color center.

[0177] Beam splitter 305, dichroic mirror 308, and filter assembly 309 are used to separate, combine, and filter light of different wavelengths, propagation directions, or polarization states.

[0178] The second polarization control element 306 is used to analyze the magneto-optical Kerr polarization change in the reflected light of the multilayer magnetic structure 502 under test.

[0179] The fifth lens 307 is used to image the collected light onto the image acquisition component 310. The image acquisition component 310 may include a CCD camera, a CMOS camera, an sCMOS camera, an intensified camera, or other photodetectors for acquiring MOKE images, NV fluorescence images, ODMR frequency scan images, or time-series images.

[0180] like Figure 3As shown, the microwave excitation module 400 is used to apply a microwave signal to the NV color center in the diamond NV sensing structure 501 to obtain the ODMR spectrum. The microwave excitation module 400 includes a GHz signal generator 401, a modulation or switching circuit 402, a power amplifier 403, a transmission and matching circuit 404, an antenna position adjustment structure 405, a microwave antenna clamping tool 406, and a microwave antenna 407.

[0181] The GHz signal generator 401 is used to generate microwave sweep signals.

[0182] The modulation or switching circuit 402 is used to control the on / off state, pulse, modulation, or frequency scanning of microwave signals.

[0183] Power amplifier 403 is used to amplify microwave signals.

[0184] The transmission and matching circuit 404 is used to transmit microwave signals to the microwave antenna 407.

[0185] The antenna position adjustment structure 405 and the microwave antenna clamping tool 406 are used to adjust and fix the position of the microwave antenna 407 relative to the diamond NV sensing structure 501.

[0186] Microwave antenna 407 is positioned near diamond NV sensing structure 501 to generate microwave magnetic field in the region where NV color center is located.

[0187] like Figure 3 As shown, the in-situ magnetic characterization platform 500 includes a diamond NV sensing structure 501, a multilayer magnetic structure under test 502, a magnetic field excitation module 503, an electrical excitation and readout module 504, a sample stage 505, and a displacement module 506.

[0188] The diamond NV sensing structure 501 includes a diamond sheet, diamond film, diamond film, or solid-state sensing structure containing NV color centers.

[0189] As an alternative, the multilayer magnetic structure 502 to be tested can be a multilayer magnetic thin film, a magnetic heterostructure, a magnetic device, or a micro / nano magnetic structure.

[0190] The diamond NV sensing structure 501 is used to detect stray magnetic fields generated by the multilayer magnetic structure under test, so that the stray magnetic field generated by the multilayer magnetic structure under test 502 can act on the NV color center.

[0191] As an optional solution, the diamond NV sensing structure 501 is positioned above, below, or to the side of the multilayer magnetic structure 502 under test. It can be in contact with, close to, or fixedly combined with the multilayer magnetic structure 502 under test, or separated by a non-magnetic spacer layer. As long as the stray field generated by the magnetic layer under test can reach the NV detection plane and the system can complete the stable registration of adjacent states, this method can be executed.

[0192] The magnetic field excitation module 503 is used to apply preprocessing excitation, layer selective writing excitation and measurement bias field to the multilayer magnetic structure 502 under test.

[0193] As an alternative, the magnetic field excitation module 503 may include an electromagnet, a Helmholtz coil, a triaxial coil, a microcoil, a permanent magnet, a yoke structure, a current line, an on-chip conductor, or a coplanar waveguide.

[0194] The electrical excitation and readout module 504 is used to apply current excitation, voltage excitation, pulse current, pulse voltage, gate voltage or electrical readout signal to the multilayer magnetic structure 502 under test. It may include a current source, voltage source, pulse source, arbitrary waveform generator, source meter, oscilloscope, lock-in amplifier, probe, probe holder or sample electrode.

[0195] The sample stage 505 is used to support the multilayer magnetic structure 502 and / or the diamond NV sensing structure 501 to be tested.

[0196] It is worth noting that the multilayer magnetic structure 502 under test and the diamond NV sensing structure 501 are relatively stationary.

[0197] The displacement module 506 is used to adjust the relative positions between the multilayer magnetic structure under test 502, the sample stage 505, the microscope objective 304 and the microwave antenna 407, so that the MOKE image and the NV magnetic field image correspond to the same area or the registration area.

[0198] like Figure 3 As shown, the data processing module 600 is connected to the MOKE light source module 100, the NV light source module 200, the common incident and collection optical path module 300, the microwave excitation module 400, and the in-situ magnetic characterization platform 500, respectively.

[0199] The data processing module 600 controls the operating states of the MOKE light source 101, laser 201, image acquisition component 310, GHz signal generator 401, magnetic field excitation module 503, electrical excitation and readout module 504, and displacement module 506. It also acquires prior information about the multilayer magnetic structure 502 under test, determines the target magnetic layer, non-target magnetic layer, and layer-selective writing excitation parameters, controls the multilayer magnetic structure 502 under test to enter a reference magnetic state, acquires a first bimodal image, controls the application of layer-selective writing excitation, and acquires a second bimodal image in a stable magnetic state after writing. The data processing module 600 further performs spatial registration, background subtraction, intensity normalization, differential processing, and layer attribution determination on the first and second bimodal images to generate layer-resolution test results.

[0200] When the system is working, the data processing module 600 determines the target magnetic layer, non-target magnetic layer, and layer selective writing excitation parameters based on the prior information of the multilayer magnetic structure 502 under test. The data processing module 600 controls the magnetic field excitation module 503 and / or the electrical excitation and readout module 504 to apply pre-processing excitation, so that the multilayer magnetic structure 502 under test enters the reference magnetic state. In the reference magnetic state, the MOKE detection light emitted by the MOKE light source module 100 is irradiated onto the multilayer magnetic structure 502 under test through the common incident and collection optical path module 300. The reflected light from the multilayer magnetic structure 502 under test is processed by the second polarization control element 306 and the image acquisition component 310 to form a first MOKE image. Simultaneously or sequentially, the excitation light emitted by the NV light source module 200 is irradiated onto the diamond NV sensing structure 501. The microwave excitation module 400 applies a microwave sweep frequency signal to the NV color center. The image acquisition component 310 acquires NV fluorescence images at different microwave frequencies, which are processed by ODMR to form a first NV magnetic field image. The first MOKE image and the first NV magnetic field image constitute the first bimodal image.

[0201] The data processing module 600 controls the magnetic field excitation module 503 and / or the electrical excitation and readout module 504 to apply layer-selective writing excitation to the multilayer magnetic structure 502 under test, causing the target magnetic layer to undergo magnetic domain nucleation, domain wall movement, magnetization reversal, or magnetization rotation, while keeping the non-target magnetic layers stable or within a preset range of variation. After the layer-selective writing excitation ends, the multilayer magnetic structure 502 under test enters a stable magnetic state after writing. In this stable magnetic state, the data processing module 600 controls the system to acquire a second MOKE image and a second NV magnetic field image, which together constitute a second bimodal image.

[0202] The data processing module 600 processes the first bimodal image and the second bimodal image to obtain the MOKE change and NV change, and generates layer-resolved test results based on the MOKE change, NV change, their spatial correspondence, and the layer-selective writing excitation conditions. The layer-resolved test results may include the target magnetic layer response, the buried magnetic layer response, the near-compensated magnetic structure response, the multilayer common response, or the interlayer correlation response.

[0203] Through the aforementioned system structure, MOKE imaging and NV magnetic field imaging can be acquired for the same test area under the same device or a registrationable device, improving the spatial consistency between the two modal images. The data processing module 600 uniformly controls the preprocessing excitation, layer-selective writing excitation, post-writing retention, dual-modal readout, and image processing processes, improving the repeatability of comparisons between adjacent magnetic states. By jointly judging the magnetic domain morphology information provided by the MOKE image and the stray field distribution information provided by the NV magnetic field image, the ability to distinguish the responses of different magnetic layers in a multilayer magnetic structure can be improved.

[0204] It is worth noting that the above system structure is only one optional implementation. Other structures that can complete MOKE image acquisition, NV magnetic field image acquisition, layer selective writing, post-write retention, image registration, and layer attribution processing can also be used.

[0205] It is worth noting that the division of modules is used to describe the functional composition of the system, and does not require that each module be an independent physical structure. In actual systems, the same optical element, mechanical structure, circuit structure or control structure can be shared by multiple modules.

[0206] In this embodiment of the invention, the collaborative operation of the MOKE light source module, NV light source module, shared incident and collection optical path module, microwave excitation module, in-situ magnetic characterization platform, and data processing module enables the acquisition of MOKE images and NV magnetic field images within the same test area or registrationable area, improving the spatial consistency between the two-modal images. By applying preprocessing excitation, layer-selective writing excitation, and measurement bias field to the multilayer magnetic structure under test through the in-situ magnetic characterization platform, and by having the data processing module uniformly control the writing, holding, reading, and image processing processes, the repeatability of adjacent magnetic state comparisons and the reliability of layer attribution determination can be improved. By jointly utilizing the magnetic domain morphology information provided by the MOKE image and the stray field distribution information provided by the NV magnetic field image, the ability to distinguish surface magnetic layers, buried magnetic layers, near-compensated magnetic structures, and interlayer correlation responses in multilayer magnetic structures can be enhanced.

[0207] The following embodiments of the present invention all perform layer-resolved testing based on comparison of adjacent magnetic states. In each write, hold, and read cycle, the previous magnetic state, target magnetic layer, non-target magnetic layer, layer-selective write excitation, and allowable variation range of the non-target magnetic layer of the multilayer magnetic structure under test are first determined. Subsequently, after the write excitation is turned off and the multilayer magnetic structure under test enters a stable magnetic state, MOKE images and NV magnetic field images of the same region or a registerable region are acquired, and layer assignment is determined based on the excitation conditions, prior information of the multilayer magnetic structure under test, and the correspondence between MOKE changes and NV changes.

[0208] If, after a single write, hold, and read cycle, the change in the non-target magnetic layer exceeds a preset allowable range, the magnetic state change corresponding to that cycle can be determined as a multilayer co-response, or the layer selective write excitation parameters can be readjusted and the test repeated. If it is necessary to track the continuous magnetization evolution process, hold-state reads can be repeated between multiple write pulses to obtain a dual-modal image sequence of multiple adjacent magnetic states.

[0209] Example 1: Taking a PMA / PMA dual-layer structure as an example, the multilayer magnetic structure under test includes a surface hard magnetic PMA layer and a buried soft magnetic PMA layer. The coercive field of the surface hard magnetic PMA layer is greater than that of the buried soft magnetic PMA layer, and both layers exhibit perpendicular magnetic anisotropy. Based on the difference in the flipping regions of the two layers, selective writing conditions can be formed by exciting the layers with magnetic fields of different amplitudes or directions.

[0210] A pretreatment magnetic field is applied to the multilayer magnetic structure under test to induce a uniform saturation state in the buried soft magnetic PMA layer along a predetermined direction. The amplitude of the pretreatment magnetic field is determined based on the saturation field of the buried soft magnetic PMA layer, thereby eliminating or weakening the local magnetic domain structure of the buried soft magnetic PMA layer. If the surface hard magnetic PMA layer undergoes a change in magnetic state during the pretreatment process, the final magnetic state of the surface hard magnetic PMA layer is taken as the known initial state for subsequent tests.

[0211] With the buried soft magnetic PMA layer kept saturated, a write excitation is applied to the multilayer magnetic structure under test to drive the surface hard magnetic PMA layer, causing magnetic domain nucleation, domain wall movement, or partial flipping in the surface hard magnetic PMA layer. The direction, amplitude, duration, or number of pulses of this write excitation are determined according to the flipping intervals of the surface hard magnetic PMA layer and the buried soft magnetic PMA layer, so that the surface hard magnetic PMA layer generates a magnetic domain response while keeping the buried soft magnetic PMA layer in a saturated state.

[0212] After the write excitation was turned off and the magnetic state stabilized, MOKE and NV magnetic field images of the same region were acquired. Since the buried soft magnetic PMA layer was in a saturated state, the magnetic domains appearing in the MOKE image mainly corresponded to the magnetic domain response of the surface hard magnetic PMA layer; the stray field characteristics in the NV magnetic field image corresponding to the spatial region of the MOKE magnetic domains could be used as a reference for the NV response of the surface hard magnetic PMA layer. This magnetic state was recorded as the reference magnetic state.

[0213] Based on the established magnetic domains of the surface hard magnetic PMA layer, a reverse layer-selective writing excitation with a low amplitude is applied. The intensity of this layer-selective writing excitation reaches the flipping range of the buried soft magnetic PMA layer, but is lower than the condition for changing the existing magnetic domain state of the surface hard magnetic PMA layer. This causes magnetic domain nucleation, domain wall movement, or partial flipping in the buried soft magnetic PMA layer, while keeping the magnetic domain position and morphology of the surface hard magnetic PMA layer within a preset range of variation.

[0214] After each write stimulus, the write stimulus is turned off and the magnetic state is allowed to stabilize before acquiring a second MOKE image and a second NV magnetic field image. The second MOKE image is compared with the first MOKE image, and the second NV magnetic field image is compared with the first NV magnetic field image. If the position and morphology of the magnetic domains in the surface hard magnetic PMA layer in the MOKE image remain stable, while the NV magnetic field image shows newly added or continuously evolving magnetic field structures relative to the reference magnetic state, then the newly added or evolved magnetic field response is attributed to the buried soft magnetic PMA layer.

[0215] This embodiment uses a three-state timing sequence of "saturation of the buried soft magnetic layer, formation of magnetic domains in the surface hard magnetic layer, and keeping the surface hard magnetic layer unchanged while driving the buried soft magnetic layer" to distinguish the responses of the surface hard magnetic PMA layer and the buried soft magnetic PMA layer in the adjacent magnetic state difference, thereby improving the layer assignment accuracy of the responses of magnetic layers at different depths in the PMA / PMA double-layer structure.

[0216] It is worth noting that Embodiment 1 of this application only uses the surface hard magnetic PMA layer and the buried soft magnetic PMA layer as examples. In practical applications, the surface soft magnetic PMA layer and the buried hard magnetic PMA layer are also applicable to this application.

[0217] Example 2: Taking a PMA / IMA structure as an example, the multilayer magnetic structure under test includes a PMA layer and an IMA layer. The PMA layer has perpendicular magnetic anisotropy, and the IMA layer has in-plane magnetic anisotropy. Since the magnetic anisotropy directions of the two layers are different, out-of-plane excitation and in-plane excitation can be used respectively to form selective writing conditions for the layers.

[0218] In one testing method, an out-of-plane excitation selectively drives the PMA layer. A pre-processing excitation is applied along the in-plane easy axis direction of the IMA layer to bring the IMA layer into a known and stable in-plane magnetization state, and to place the PMA layer in an initial magnetic state suitable for subsequent out-of-plane writing. After removing the pre-processing excitation, or retaining a measurement bias field below the magnetization change threshold, a first MOKE image and a first NV magnetic field image are acquired.

[0219] A layer-selective write excitation is applied along the out-of-plane direction, causing magnetic domain nucleation, domain wall movement, or magnetization reversal in the PMA layer, while the IMA layer maintains its original in-plane magnetization state or remains within a preset variation range. After the write excitation is turned off and the magnetic state stabilizes, a second MOKE image and a second NV magnetic field image are acquired. If a magnetic domain change matching the magnetization component of the PMA layer appears in the poloidal MOKE image, and a stray field change corresponding to the magnetic domain change space appears in the NV magnetic field image, while the IMA layer remains stable, then the change is attributed to the PMA layer.

[0220] In another testing method, an in-plane excitation selectively drives the IMA layer. First, the PMA layer is brought into a stable out-of-plane saturation state, a stable magnetic domain state, or another magnetic state that remains stable under subsequent in-plane excitation, and a first MOKE image and a first NV magnetic field image are acquired. Then, an in-plane layer selective write excitation is applied along the easy axis of the IMA layer, causing the IMA layer to undergo magnetization rotation, domain wall movement, or flipping, while keeping the PMA layer stable or within a preset range of variation. After the write excitation is turned off and the magnetic state stabilizes, a second MOKE image and a second NV magnetic field image are acquired.

[0221] If the magnetic domains of the PMA layer remain stable as shown in the poloidal MOKE image, while a new magnetic field change appears in the NV magnetic field image, then this change is attributed to the IMA layer based on the in-plane excitation conditions. If the system is configured with longitudinal or transverse MOKE, the longitudinal or transverse MOKE signal can also be used to assist in confirming the in-plane magnetization change of the IMA layer. If both the PMA and IMA layers exhibit changes exceeding the preset range, the results are determined as a multilayer common response or an interlayer correlated response based on the MOKE change, the NV change, and their spatial correspondence, or the excitation window is readjusted before proceeding to the next round of testing.

[0222] It is worth noting that the location of the PMA layer and the IMA layer is not limited in the second embodiment of this application.

[0223] In this second embodiment, Bx, By, and Bz are used to describe the total stray field vector at the NV detection plane and provide constraints for quantitative analysis. The attribution of the PMA layer and IMA layer is determined based on the magnetic anisotropy direction, excitation direction selectivity, stability of non-target magnetic layers, and the relationship of dual-modal image changes, thereby improving the reliability of distinguishing different magnetic anisotropy layers in the PMA / IMA structure.

[0224] Example 3: Taking a top SAF and buried magnetic layer structure as an example, the multilayer magnetic structure under test includes a top near-compensated SAF structure and a buried magnetic layer below it. The top SAF structure includes at least two magnetic sublayers with opposite magnetic moment directions. The far-field stray fields generated by the two magnetic sublayers in the near-antiparallel compensation state can cancel each other out, resulting in a weaker net stray field response at the NV detection plane. Simultaneously, the top SAF structure is located in the near-surface region, and its magnetic domain or magnetization changes can be obtained with a high response through MOKE imaging. The buried magnetic layer is located below the top SAF structure, and its magneto-optical signal may be affected by the upper film stack, while its stray field changes can be detected by the diamond NV sensing structure.

[0225] In one testing method, a buried magnetic layer is driven while maintaining the stability of the top SAF structure. First, a reference magnetic state is established through preprocessing excitation, placing the top SAF structure in a known near-compensated state and the buried magnetic layer in a predetermined initial magnetic state. Then, a first MOKE image and a first NV magnetic field image are acquired. Next, a first set of layer-selective write excitations is applied, causing magnetic domain nucleation, domain wall movement, magnetization flipping, or magnetization rotation in the buried magnetic layer, while simultaneously maintaining the magnetic domain state and near-compensated state of the top SAF structure stable or within a preset variation range. After the write excitations are turned off and the magnetic state stabilizes, a second MOKE image and a second NV magnetic field image are acquired. If the magnetic domains of the top SAF structure remain essentially stable in the MOKE image, while new or evolved magnetic field structures appear in the NV magnetic field image, then the NV magnetic field change is attributed to the buried magnetic layer.

[0226] In another testing method, the top SAF structure is driven while the buried magnetic layer remains stable. First, the buried magnetic layer is brought into a magnetic state that remains stable under subsequent excitation, and a first MOKE image and a first NV magnetic field image are acquired. Then, a second set of layer-selective write excitations is applied, causing magnetic domain nucleation, domain wall movement, or magnetization reversal in the top SAF structure, while maintaining the buried magnetic layer stable. After the write excitations are turned off and the magnetic state stabilizes, a second MOKE image and a second NV magnetic field image are acquired. If the MOKE image shows changes in the magnetic domains of the top SAF structure, and the net stray field change in the NV magnetic field image is lower than a preset NV change threshold, or significantly weaker than the NV change corresponding to a normal uncompensated magnetic layer, then the response is attributed to the top SAF structure, based on the near-compensation characteristics of the top SAF structure.

[0227] During multiple write, hold, and read operations, the presence of synchronous, stepwise, or interlayer correlated responses between the top SAF structure and the buried magnetic layer can be determined based on the order of occurrence, spatial correspondence, and excitation threshold relationship of the MOKE changes in the top SAF structure and the NV magnetic field changes in the buried magnetic layer. This embodiment improves the response differentiation capability between the top near-compensated SAF structure and the buried magnetic layer by using the differential criterion between the near-surface MOKE response and the net NV stray field response.

[0228] Example 3: Taking NV measurement under a thick capping layer or opaque structure as an example, the magnetic layer under test is covered by a thick capping layer, encapsulation layer, metal electrodes, metal interconnects, on-chip wires, or other opaque structures. The capping structure can cause the MOKE illumination light or reflected light to be absorbed, scattered, or blocked. In this scenario, the diamond NV sensing structure is placed outside the capping structure, so that the stray magnetic field generated by the covered magnetic layer can act on the NV color center in the diamond NV sensing structure.

[0229] A pre-processing excitation, formed by a magnetic field, current, voltage, or a combination thereof, induces the covered magnetic layer into a known reference magnetic state. Under this reference magnetic state, diamond NV centers are excited and microwave frequency sweeps are performed to acquire a first NV fluorescence image sequence, which is then processed by ODMR to generate a first NV magnetic field image. Subsequently, magnetic field pulses, current pulses, voltage pulses, or other write excitations are applied to the covered magnetic layer, causing magnetic domain nucleation, domain wall movement, magnetization reversal, or magnetization rotation. After the write excitation is turned off and the magnetic state stabilizes, a second NV fluorescence image sequence is acquired in the same region, generating a second NV magnetic field image.

[0230] Background subtraction, spatial registration, differential processing, and magnetic field structure comparison or time series comparison are performed on the first and second NV magnetic field images to obtain the NV magnetic field changes of the covered magnetic layer before and after writing. If only one responsive magnetic layer exists under the covered structure, the magnetic response of that magnetic layer can be determined based on the NV magnetic field change. If multiple potentially responsive magnetic layers exist under the covered structure, layer attribution is determined by combining prior sample information, layer-selective writing excitation conditions, and adjacent state change sequences.

[0231] This fourth embodiment places a diamond NV sensing structure on the outside of the covered structure and uses NV magnetic field images to compare the stray field changes before and after writing the covered magnetic layer, enabling the magnetic response in the covered structure to be characterized through stray field imaging. This embodiment can be used as a supplement to the dual-modal layer-resolved testing method, or combined with dual-modal testing results from the MOKE acquisition area to improve the magnetic response analysis capability in complex device structures.

[0232] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer device, specifically, a computer device can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0233] This invention provides a computer device, including a memory and a processor. The memory stores information including program instructions, and the processor controls the execution of the program instructions. When the program instructions are loaded and executed by the processor, they implement the steps of the above-described embodiment of the multilayer magnetic structure layer resolution test method based on dual-modal images. For a detailed description, please refer to the above-described embodiment of the multilayer magnetic structure layer resolution test method based on dual-modal images.

[0234] The following is for reference. Figure 4 It shows a schematic diagram of the structure of a computer device 700 suitable for implementing the embodiments of this application.

[0235] like Figure 4 As shown, the computer device 700 includes a central processing unit (CPU) 701, which can perform various appropriate tasks and processes based on programs stored in read-only memory (ROM) 702 or programs loaded from storage section 708 into random access memory (RAM) 703. The RAM 703 also stores various programs and data required for the operation of the computer device 700. The CPU 701, ROM 702, and RAM 703 are interconnected via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0236] The following components are connected to the I / O interface 705: an input section 706 including a keyboard, mouse, etc.; an output section 707 including a cathode ray tube (CRT), liquid crystal feedback (LCD), etc., and speakers, etc.; a storage section 708 including a hard disk, etc.; and a communication section 709 including a network interface card such as a LAN card, modem, etc. The communication section 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the I / O interface 705 as needed. A removable medium 711, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 710 as needed so that computer programs read from it can be installed in the storage section 708 as needed.

[0237] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program including program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 709, and / or installed from removable medium 711.

[0238] This invention also discloses a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the method described above. Computer-readable media include permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient media, such as modulated data signals and carrier waves.

[0239] For ease of description, the above devices are described separately by function as various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0240] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0241] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0242] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0243] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0244] The acquisition, storage, use, and processing of data in this application all comply with the relevant provisions of national laws and regulations.

[0245] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0246] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0247] This application can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0248] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0249] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for resolving multilayer magnetic structures based on dual-modal images, characterized in that, The method includes: Acquire prior information about the multilayer magnetic structure under test, and determine the target magnetic layer, non-target magnetic layer, and layer selective writing excitation based on the prior information; Based on the target magnetic layer and the non-target magnetic layer, a reference magnetic state is constructed, and a first dual-mode image of the multilayer magnetic structure under test is acquired under the reference magnetic state. Based on the target magnetic layer and non-target magnetic layer, the layer selective writing excitation is applied, and the second bimodal image of the multilayer magnetic structure under test is acquired in a stable magnetic state. Layer resolution processing is performed on the first bimodal image and the second bimodal image to generate layer resolution test results.

2. The method for multilayer magnetic structure resolution testing based on dual-modal images according to claim 1, characterized in that, The construction of a reference magnetic state based on the target magnetic layer and the non-target magnetic layer includes: A pre-processing excitation is applied to the multilayer magnetic structure under test, so that the non-target magnetic layer is in a stable magnetic state and the target magnetic layer is in an initial magnetic state.

3. The method for multilayer magnetic structure resolution testing based on dual-modal images according to claim 1, characterized in that, The first bimodal image includes a first MOKE image and a first NV magnetic field image; The acquisition of the first dual-mode image of the multilayer magnetic structure under test in the reference magnetic state includes: The MOKE pattern is determined based on the prior information, and wide-field illumination is applied to the multilayer magnetic structure under test based on the MOKE pattern to acquire the first MOKE image. A predetermined measurement bias field is applied to the multilayer magnetic structure under test, and diamond NV color centers are excited to acquire a preliminary ODMR spectrum. Based on the prepared ODMR spectrum, determine whether the eight ODMR resonance peaks overlap; If so, apply in-plane bias fields along the horizontal and vertical axes, acquire reference ODMR spectra, and calibrate NV axis pairs and crystal orientation relationships based on the reference ODMR spectra; In the reference magnetic state, an in-plane bias field and a perpendicular bias field are applied to the multilayer magnetic structure under test, the ODMR spectrum under test is acquired, and a first NV magnetic field image is generated based on the ODMR spectrum under test, the calibrated NV axis pairs and crystal orientation relationship. If not, perform eight-peak fitting and projection transformation on the prepared ODMR spectrum at each pixel location to obtain the first NV magnetic field image.

4. The method for multilayer magnetic structure resolution testing based on dual-modal images according to claim 1, characterized in that, The step of applying the layer-selective write stimulus based on the target magnetic layer and the non-target magnetic layer includes: A layer-selective writing excitation is applied to the multilayer magnetic structure under test, so that the non-target magnetic layer is within a preset variation range, causing the target magnetic layer to undergo a change in magnetization state.

5. The method for multilayer magnetic structure resolution testing based on dual-modal images according to claim 1, characterized in that, The step of performing layer resolution processing based on the first bimodal image and the second bimodal image to generate layer resolution test results includes: Image preprocessing is performed on the first bimodal image and the second bimodal image to obtain a preprocessed bimodal image pair. The image preprocessing includes spatial registration, background subtraction, and intensity normalization. The dual-modal image pairs are subjected to differential processing to obtain the image changes between adjacent states; Layer assignment is performed based on the image changes of the adjacent states to obtain the layer resolution test results.

6. The method for multilayer magnetic structure resolution testing based on dual-modal images according to claim 5, characterized in that, The image changes include MOKE changes and NV changes; The step of performing layer assignment processing based on the image changes of adjacent states to obtain the layer resolution test results includes: The layer attribution result is generated by comparing the MOKE change and the NV change. Using the current magnetic state as a reference magnetic state, the layer selectively writes the update stimulus, and continues to execute the step of acquiring the first bimodal image of the multilayer magnetic structure under test in the reference magnetic state; Based on the layer attribution results obtained from multiple updates, layer resolution test results are generated.

7. The method for multilayer magnetic structure resolution testing based on dual-modal images according to claim 6, characterized in that, The step of comparing the MOKE change and the NV change to generate layer attribution results includes: If the MOKE change is greater than a preset MOKE change threshold, the NV change is greater than a preset NV change threshold, and the overlap rate between the magnetic domain region corresponding to the MOKE change and the magnetic domain region corresponding to the NV change is greater than a preset overlap rate threshold, the target magnetic layer is determined as the layer assignment result. If the MOKE change is greater than a preset MOKE change threshold, the NV change is greater than a preset NV change threshold, and the overlap rate between the domain region corresponding to the MOKE change and the domain region corresponding to the NV change is less than or equal to the overlap rate threshold, the interlayer coupling response is determined as the layer attribution result. If the MOKE change is greater than the preset MOKE change threshold, and the NV change is less than or equal to the preset NV change threshold, the near-compensated magnetic structure is determined as the layer assignment result. If the MOKE change is less than or equal to a preset MOKE change threshold, and the NV change is greater than a preset NV change threshold, the buried magnetic layer is determined as the layer assignment result.

8. A multilayer magnetic structure layer resolution testing system based on dual-modal imaging, characterized in that, The system applied to the multilayer magnetic structure layer resolution testing method based on dual-modal images according to any one of claims 1 to 7, the system comprising: The MOKE light source module is used to provide MOKE detection light to the multilayer magnetic structure under test. NV light source module, used to excite the NV color centers of diamond to produce fluorescence; A shared incident and collection optical path module is used to guide the MOKE detection light and / or the excitation light emitted by the NV light source module to the diamond NV sensing structure of the multilayer magnetic structure under test and the in-situ magnetic characterization platform, and to collect the reflected light of the multilayer magnetic structure under test and the fluorescence generated by the diamond NV color center. A microwave excitation module is used to apply a microwave sweep signal to the diamond NV color center to obtain the ODMR spectrum. The in-situ magnetic characterization platform includes a diamond NV sensing structure, a multilayer magnetic structure under test, a magnetic field excitation module, an electrical excitation and readout module, a sample stage, and a displacement module. The diamond NV sensing structure is disposed on the multilayer magnetic structure under test and is used to detect the stray magnetic field generated by the multilayer magnetic structure under test. The magnetic field excitation module is used to apply preprocessing excitation, layer selective writing excitation and measurement bias field to the multilayer magnetic structure under test; The electrical excitation and readout module is used to apply current excitation, voltage excitation, or electrical readout signal to the multilayer magnetic structure under test. The sample stage is used to support the multilayer magnetic structure to be tested and / or the diamond NV sensing structure. The displacement module is used to adjust the relative position between the multilayer magnetic structure under test and the diamond NV sensing structure, and to make the first bimodal image and the second bimodal image correspond to the same area. The data processing module is connected to the MOKE light source module, NV light source module, common incident and collection optical path module, microwave excitation module, and in-situ magnetic characterization platform, respectively. It is used to control the multilayer magnetic structure under test to enter the reference magnetic state; acquire a first bimodal image; apply the layer selective writing excitation and acquire a second bimodal image under stable magnetic state; and perform layer resolution processing based on the first and second bimodal images to generate layer resolution test results.

9. A computer device comprising a memory and a processor, the memory for storing information including program instructions, and the processor for controlling the execution of the program instructions, characterized in that, When the program instructions are loaded and executed by the processor, they implement the multi-layer magnetic structure layer resolution test method based on dual-modal images as described in any one of claims 1 to 7.

10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the multi-layer magnetic structure layer resolution test method based on dual-modal images as described in any one of claims 1 to 7.

Images