Three-dimensional reconstruction method of porous Ni-YSZ electrodes for solid oxide batteries

Abstract

The present invention provides a three-dimensional reconstruction method for porous Ni-YSZ electrodes of solid oxide batteries, comprising the following steps: S1, vacuum cold mounting; S2, argon ion polishing; S3, sample pretreatment and fixation; S4, FIB continuous sectioning and In-lens and SE2 dual-channel image acquisition; S5, deep learning analysis. The beneficial effects of the present invention are: providing a three-dimensional reconstruction method for porous Ni-YSZ electrodes of solid oxide batteries, by integrating in-situ characterization and intelligent image analysis technology, achieving accurate analysis of multiphase material structure, with the advantages of high sample preparation stability, good three-dimensional reconstruction accuracy, and high degree of phase identification automation. It is particularly suitable for the study of the microstructural evolution of the electrode / electrolyte interface of solid oxide fuel cells, and is of great value in revealing the material performance degradation mechanism under high-temperature and complex working conditions.

Description

Technical Field

[0001] The present invention relates to a battery, and in particular to a three-dimensional reconstruction method for a porous Ni-YSZ electrode of a solid oxide battery. Background Art

[0002] Solid oxide fuel cells / electrolyzers (SOFCs / SOECs) are highly efficient and reversible energy conversion devices. Their performance and lifespan critically depend on the microstructural properties of their electrode materials, particularly the structural stability of the fuel electrode during water electrolysis to produce hydrogen in the SOEC mode. Currently widely used Ni-YSZ cermet electrodes face significant challenges in high-temperature steam electrolysis environments: Ni particle migration, agglomeration, and coarsening not only reduce the length of the three-phase interfaces (TPBs) but also exacerbate the thermal expansion mismatch with the YSZ electrolyte, leading to rapid degradation of electrode performance. While FIB serial section tomography (FIB) offers the potential for three-dimensional microstructural reconstruction with nanometer-scale resolution, its application is severely limited by the quality of material contrast. Due to the similar average atomic numbers of Ni and YSZ, conventional backscattered electron imaging struggles to provide sufficient interphase contrast. Furthermore, intensity variations caused by differences in Ni grain orientation present significant technical obstacles for image segmentation and three-dimensional reconstruction. In the existing technology, although the contrast can be improved to a certain extent by optimizing SEM imaging parameters, it still cannot meet the requirements of high-precision three-dimensional characterization of the Ni-YSZ system. This seriously restricts the in-depth understanding of the evolution of electrode microstructure and the mechanism of electrochemical performance degradation, and becomes a key technical bottleneck in the development of long-life SOEC systems.

[0003] In the existing technology, CN115830226A proposes a three-dimensional porous media reconstruction method based on FIB-SEM slicing and deep learning. Through grayscale statistical analysis, binary slicing processing and deep learning dimension expansion, high-precision three-dimensional pore structure modeling is achieved, and thermal conductivity is further predicted by combining pore parameter calibration and regression analysis.

[0004] The FIB-SEM slice reconstruction method proposed in CN115830226A is not limited by the inherent lack of contrast in the material itself and is relatively easy to use. However, it relies heavily on simulation reconstruction rather than experimental reconstruction. Therefore, the accuracy of its reconstructed structure lacks credibility.

[0005] CN114518378A proposes a FIB-SEM 3D reconstruction method that enhances sample conductivity using a metal conductive mesh. The core steps include: depositing a Pt protective layer, nanomanipulator-assisted sample extraction, and welding the sample to a metal conductive mesh (copper / nickel / molybdenum mesh) to form a three-sided conductive structure. This method uses the conductive mesh to extract accumulated charge, significantly reducing image drift caused by charging effects and improving the 3D reconstruction quality of non-conductive samples.

[0006] The FIB-SEM slice reconstruction method proposed in CN114518378A requires a nanomanipulator to assist in sample extraction, a complex and time-consuming operation that is inefficient in practical applications. The method of extracting charge through a conductive mesh is similar in principle to existing "gold / carbon coating" or "ion beam-assisted conduction" solutions, which are unnecessary and may introduce additional stress due to welding process issues.

[0007] CN114723878A proposes a spherical particle-based FIB-SEM 3D reconstruction correction method. This method uses spherical particles attached to the material surface as a geometric calibration reference, combined with epoxy resin pore filling and ion beam polishing, to correct for distortion in the reconstructed image. Its core innovation lies in leveraging the known size and shape of the spherical particles (on the same order of magnitude as the material's microstructure) to correct for image distortion caused by ion beam tilt (54°) and SEM drift.

[0008] The FIB-SEM three-dimensional reconstruction correction method proposed in CN114723878A is only applicable to the reconstruction of spherical particles, and has limited correction effect on planar or irregular structures (such as layered materials). Planar samples can correct the image sequence through known angle differences and alignment algorithms.

[0009] In summary, conventional FIB-SEM reconstruction techniques have the following core limitations when applied to the Ni / YSZ system: First, due to the small difference in the average atomic numbers of Ni and YSZ, backscattered electron imaging (BSE) cannot effectively distinguish the two phases using traditional grayscale threshold segmentation methods. Second, the preferred orientation of Ni grains can cause fluctuations in the electron backscatter diffraction signal intensity, resulting in false contrast within the same phase, seriously interfering with the reliability of automatic segmentation algorithms. Furthermore, existing drift mitigation methods (such as metal mesh welding or mechanical clamping) are extremely complex and pose a risk of sample damage. Furthermore, when full experimental reconstruction is unable to obtain high-contrast images, manual annotation of numerous slices is required, regardless of whether machine learning is employed. Furthermore, due to the subjectivity of manual annotation, the cross-validation accuracy of the pore / Ni / YSZ three-phase segmentation is often poor, making it difficult to meet the requirements of submicron interface curvature analysis. These technical limitations make it impossible to accurately correlate the structure-activity relationship between electrode microstructural evolution and electrochemical performance degradation, severely restricting research on performance optimization and life prediction of SOEC electrode materials. Summary of the Invention

[0010] In order to solve the problems in the prior art, the present invention provides a three-dimensional reconstruction method for a porous Ni-YSZ electrode of a solid oxide battery.

[0011] The present invention provides a three-dimensional reconstruction method for a porous Ni-YSZ electrode of a solid oxide battery, comprising the following steps:

[0012] S1, vacuum cold mounting;

[0013] S2, argon ion polishing;

[0014] S3, sample pretreatment and fixation;

[0015] S4, FIB serial sectioning and In-lens and SE2 dual-channel image acquisition;

[0016] S5. Deep learning analysis.

[0017] As a further improvement of the present invention, step S5 includes: training a U-net neural network model, verifying whether the accuracy meets the expectations, and if not, performing local manual correction. If it meets the expectations, applying the model to all slices, and then reconstructing the 3D geometry, wherein the overall verification accuracy is >85%, the verification accuracy of the easily identifiable phase (YSZ\pore) is >90%, and the verification accuracy of the difficult-to-identify phase (Ni) is >80%.

[0018] As a further improvement of the present invention, step S1 includes: cold mounting the sample using epoxy resin under vacuum.

[0019] As a further improvement of the present invention, step S1 includes: cold mounting the sample using Struers Epofix epoxy resin in a vacuum chamber (pressure ≤10⁻³ Pa) for 8-10 hours.

[0020] As a further improvement of the present invention, step S2 includes: using an argon ion polisher to polish the sample to a flat cross section and thin the surface electrolyte.

[0021] As a further improvement of the present invention, step S2 includes: after mounting, the sample is polished with 1200-grit sandpaper until the LSCF-CGO layer is substantially removed, revealing a bright electrolyte layer; then, the cross section and plane are polished for 2 hours and 20 minutes, respectively, using a JEOL IB-19530 argon ion polisher (accelerating voltage 5.5 kV) to flatten the cross section and substantially remove the surface electrolyte; and then, a Pt conductive film protective layer approximately 10 nm thick is plated on the surface of the cell to prevent charge accumulation.

[0022] As a further improvement of the present invention, step S3 includes the following sub-steps:

[0023] S301, plating a Pt film on the sample surface;

[0024] S302, fixing the sample to a nail table with a pressing sheet;

[0025] S303, depositing a Pt protective layer in the target area;

[0026] S304: Processing grooves on both sides of the target area.

[0027] As a further improvement of the present invention, step S303 includes: sending the sample to a dual-beam electron microscope for processing and observation, using ion beam induced deposition technology to deposit a Pt protective layer on the surface of the target area, tilting the stage, adjusting the working distance to the concentric point and then performing cross-section processing, using a Ga+ ion beam, and using tilt compensation and dynamic focusing functions to correct the geometric distortion of the image to ensure the quality of the SEM image.

[0028] As a further improvement of the present invention, a 1 μm thick Pt protective layer is again deposited on the surface of the target area using ion beam induced deposition technology.

[0029] As a further improvement of the present invention, cross-section processing was performed after the FIB stage was tilted 54° and adjusted to the concentric point at a working distance of 5.1 mm.

[0030] As a further improvement of the present invention, step S304 includes: using a higher ion beam current to cut large grooves around the target area to avoid detector shadows during SEM imaging, and then reducing the ion beam current to a set value to perform fine processing of the final cross section.

[0031] As a further improvement of the present invention, step S4 includes: before FIB continuous slicing, calling out the TEM probe to press the sample edge to ensure a smaller displacement error during FIB slicing; then performing continuous slicing with a set layer thickness, and simultaneously performing In-lens and SE2 dual-channel image acquisition, collecting surface morphology images through the In-lens detector, and simultaneously obtaining high-contrast phase distribution images with the SE2 detector.

[0032] As a further improvement of the present invention, step S5 includes: using the last high-contrast In-lens image in step S4 as a deep learning sample, extracting the sample multi-phase feature data based on threshold segmentation and generating a label data set corresponding to the SE2 image, deep learning model training and automatic segmentation of the SE2 image sequence, local area manual optimization and model iterative training, and reconstructing 3D geometry.

[0033] As a further improvement of the present invention, step S5 includes: after the continuous sectioning is completed, pausing for a few seconds on the last image, utilizing the contrast change during the charging effect, and collecting again after a brief accumulation of charge to obtain a single In-lens image with high contrast although contrast-inverted; using the collected single In-lens image as a training sample, extracting the three-phase contrast thresholds of Ni, YSZ, and pores through multi-threshold segmentation and generating binary labels, using a U-Net architecture with an input layer size of 96×96, setting the patch size to 128, extracting 4 patches from each image, training with the Dice coefficient + cross entropy as the loss function, and finally importing the segmented SE2 sequence into 3D reconstruction software for voxel reconstruction.

[0034] The present invention provides a method for 3D reconstruction of porous Ni-YSZ electrodes in solid oxide fuel cells. By integrating in-situ characterization with intelligent image analysis, it enables precise analysis of multiphase material structures. This method offers advantages such as high sample preparation stability, excellent 3D reconstruction accuracy, and a high degree of automated phase identification. The method is particularly suitable for studying the microstructural evolution of the electrode / electrolyte interface in solid oxide fuel cells and is of great value in revealing the mechanisms of material performance degradation under high-temperature and complex operating conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other solutions can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 This is a flow chart of a three-dimensional reconstruction method for a porous Ni-YSZ electrode of a solid oxide battery according to the present invention;

[0037] Figure 2 Schematic diagram of a stage device with a tablet press;

[0038] Figure 3 This is a schematic diagram of the TEM probe limit;

[0039] Figure 4 This is the last slice image under continuous slice dual-channel sampling;

[0040] Figure 5 It is the last slice image of the In-lens after stopping slicing and re-sampling;

[0041] Figure 6 This is the effect (U-Net architecture) of using the high-contrast last slice of the In-lens to train the corresponding SE2 image. DETAILED DESCRIPTION

[0042] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0043] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the scope of protection of the present invention. In addition, the terms "first", "second" and the like are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, features defined as "first", "second" and the like may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0044] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood broadly. For example, they may refer to fixed connections, detachable connections, or integral connections; they may refer to direct connections, indirect connections through an intermediate medium, or internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0045] The present invention will be further described below with reference to the accompanying drawings and specific implementation methods.

[0046] A three-dimensional reconstruction method for porous Ni-YSZ electrodes in solid oxide batteries involves a composition contrast optimization and intelligent segmentation method for Ni-YSZ composite materials based on FIB continuous slicing technology. This method belongs to the technical field of material microstructure characterization and aims to solve technical problems such as sample drift, insufficient composition contrast, and poor phase segmentation accuracy during the FIB slicing process in traditional three-dimensional characterization.

[0047] The novel sample preparation process designed in this invention sequentially includes cold mounting with epoxy resin under vacuum, cross-section polishing and surface electrolyte thinning using an argon ion polisher, Pt protective layer deposition, sample fixation with a press pin (using a TEM probe to hold the sample firmly in place before FIB serial sectioning to prevent drift), local micromachining and FIB serial sectioning, In-lens / Se2 dual-channel acquisition (with the final high-contrast In-lens image serving as the deep learning sample), threshold-based segmentation to extract multiphase feature data from the sample and generate a labeled dataset corresponding to the Se2 image, deep learning model training and automated segmentation of the Se2 image sequence, manual optimization of local regions, and iterative model training. This sample preparation method, integrating in situ characterization with intelligent image analysis techniques, enables precise analysis of multiphase material structures. It offers advantages such as high sample preparation stability, excellent 3D reconstruction accuracy, and a high degree of automated phase identification. It is particularly suitable for studying the microstructural evolution of the electrode / electrolyte interface in solid oxide fuel cells and holds significant value in revealing the mechanisms of material performance degradation under high-temperature and complex operating conditions.

[0048] like Figure 1 As shown, the main process of the present invention includes: vacuum cold mounting, argon ion polishing, sample pretreatment and fixation, FIB-SEM dual-channel acquisition, and deep learning analysis. The specific process is as follows:

[0049] First, the sample was cold-mounted in a vacuum chamber (pressure ≤10⁻³ Pa) using Struers Epofix epoxy resin for 8-10 hours. After mounting, the sample was polished with 1200-grit sandpaper until the LSCF-CGO layer was basically removed, revealing a bright electrolyte layer. The cross-section and plane were then polished using a JEOL IB-19530 argon ion polisher (accelerating voltage 5.5 kV) for 2 hours and 20 minutes, respectively, to flatten the cross-section and basically remove the surface electrolyte. A Pt conductive film protective layer with a thickness of approximately 10 nm was then plated on the surface of the cell to prevent charge accumulation.

[0050] Then, if Figure 2As shown, a stage assembly 4 with a pressure plate is provided, and a pin stage 2 with a pressure plate 3 is used to hold the sample 1 in place. The stage assembly 4 prevents sample drift during sectioning. The sample is then transferred to a Zeiss Crossbeam 350 dual-beam electron microscope for processing and observation. Ion beam-induced deposition is used to deposit a 1 μm-thick Pt protective layer on the target area. In the FIB-SEM, the stage is tilted 54° and adjusted to the concentric point at a working distance of 5.1 mm before cross-sectioning. A 30 kV Ga+ ion beam is used. To ensure SEM image quality, tilt compensation and dynamic focusing are used to correct geometric distortion. During processing, a high ion beam current (up to 65 nA) is initially used to cut a large trench around the target area to avoid detector shadowing during SEM imaging. The beam current is then reduced to 3 nA for fine processing of the final cross-section.

[0051] Before FIB serial sectioning, Figure 3 As shown in the figure, the TEM probe was called out to press the edge of the sample to ensure a small displacement error during FIB slicing, and the TEM probe was limited to prevent sample drift during slicing. Continuous slicing was then performed with a layer thickness of 50 nm, and In-lens and SE2 dual-channel image acquisition was performed simultaneously (accelerating voltage 1.5 kV, working distance 5.1 mm). The surface morphology image was collected using the In-lens detector, and the SE2 detector was used to obtain high-contrast phase distribution images with a resolution of ≥2048×1536 pixels.

[0052] In order to solve the problem that the backscattered electron detector is difficult to achieve effective phase separation under various conditions due to the close average atomic numbers of Ni / YSZ calculated by the electron fraction weighting method and the intensity variation caused by the difference in Ni grain orientation (such as Figure 4 ), after the continuous sectioning is completed, pause for a few seconds on the last image, and use the contrast change during the charging effect. After a short period of charge accumulation, the image can be collected again to obtain a single In-lens image with high contrast despite the contrast inversion (such as Figure 5 ), using the final in-lens image as the training sample, multi-threshold segmentation was used to extract the Ni, YSZ, and pore phase contrast thresholds and generate binary labels. A U-Net architecture with an input layer size of 96×96 (including 4 layers of downsampling and upsampling) was used, with a patch size of 128. Four patches were extracted from each image, and training was performed using the Dice coefficient + cross entropy loss function (3 images per batch, batch size = 12). Given the difficulty in obtaining label images corresponding to SE2 images, verification accuracy of >90% for the YSZ and pore phases, >80% for the Ni phase, and >85% for the overall verification accuracy were achieved using only 3 sample images and approximately 500 iterations. Figure 6Finally, the segmented SE2 sequence can be imported into 3D reconstruction software such as Avizo for voxel reconstruction, and parameters such as the volume fraction of each phase and interface curvature can be quantitatively calculated to correlate material properties.

[0053] This scheme achieves non-destructive sample preparation through the combination of vacuum cold mounting and argon ion polishing, adopts dual fixation of pressing nail stage + TEM probe to solve the problem of nano-slice drift, innovatively uses charging effect to obtain high-contrast In-lens images as training samples, and adopts appropriate U-Net neural network parameters to achieve efficient training of small samples, effectively overcoming the phase segmentation difficulties caused by the close average atomic number of Ni / YSZ in traditional backscattered electron images. Ultimately, a standardized solution for the entire process of three-dimensional reconstruction from precise sample preparation, dual-channel imaging to intelligent segmentation is formed, providing a high-precision analysis method for the interface evolution study of multiphase materials.

[0054] Compared with other existing technologies, the advantages of the present invention are as follows:

[0055] (1) Compared with the traditional FIB-SEM three-dimensional reconstruction method, the present invention achieves non-destructive sample preparation by combining vacuum cold mounting with argon ion polishing, thus avoiding microstructure distortion caused by mechanical damage.

[0056] (2) The dual fixation technology of the pressing nail stage + TEM probe is used to effectively solve the problem of sample drift during the nanoscale continuous sectioning process, greatly reduce the difficulty of subsequent image sequence alignment, and significantly improve the accuracy of three-dimensional reconstruction.

[0057] (3) The charging effect is innovatively used to obtain high-contrast In-lens images as training samples, which solves the problem of insufficient contrast of traditional backscattered electron images caused by the similar average atomic numbers of Ni / YSZ, and overcomes the interference of contrast changes caused by differences in Ni grain orientation on phase segmentation.

[0058] (4) Using the U-Net neural network structure, only three training samples are needed to achieve a segmentation accuracy of more than 90% for YSZ and pore phases and more than 80% for Ni phases, which significantly reduces the demand for deep learning on the number of samples and manual labeling, and largely avoids the reduction in phase segmentation accuracy caused by the randomness of manual labeling.

[0059] (5) It provides a standardized 3D reconstruction process from precise sample preparation, dual-channel imaging to intelligent segmentation.

[0060] (6) The charging effect enhanced In-lens imaging technology and deep learning algorithm effectively solved the image segmentation problem caused by insufficient interphase contrast difference in the Ni / YSZ system.

[0061] (7) The sample fixing technology that cooperates with the TEM probe to ensure the stability of nanoscale continuous sectioning.

[0062] (8) Specific parameters of the argon ion polishing process and the FIB parameters of the micromachining process during sample preparation.

[0063] The above is a further detailed description of the present invention in conjunction with specific preferred embodiments, and the specific implementation of the present invention should not be considered to be limited to these descriptions. For those skilled in the art to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, and all of these should be considered to fall within the scope of protection of the present invention.

Claims

1. A three-dimensional reconstruction method for porous Ni-YSZ electrodes of solid oxide batteries, characterized in that: The following steps are involved: S1, vacuum cold mounting; S2, argon ion polishing; S3, sample pretreatment and fixation; S4, FIB serial sectioning and In-lens and SE2 dual-channel image acquisition; S5, deep learning analysis; Step S4 includes: before FIB continuous sectioning, calling up the TEM probe to press the sample edge to ensure a small displacement error during FIB sectioning; then performing continuous sectioning at a set layer thickness, and simultaneously performing In-lens and SE2 dual-channel image acquisition, using the In-lens detector to acquire a surface topography image, while simultaneously using the SE2 detector to acquire a high-contrast phase distribution image; Step S5 includes: after the completion of continuous slicing, pausing for a few seconds on the last image, utilizing the contrast change during the charging effect, and acquiring again after a brief accumulation of charge to obtain a single In-lens image that has been contrast-inverted but has high contrast; using the acquired single In-lens image as a training sample, extracting the three-phase contrast thresholds of Ni, YSZ, and pores through multi-threshold segmentation and generating binary labels, using a U-Net architecture with an input layer size of 96×96, setting the patch size to 128, extracting 4 patches from each image, and training with the Dice coefficient + cross entropy as the loss function. Finally, the segmented SE2 sequence is imported into the 3D reconstruction software for voxel reconstruction.

2. The method for three-dimensional reconstruction of porous Ni-YSZ electrodes for solid oxide batteries according to claim 1, characterized in that: Step S1 includes: cold mounting the sample using epoxy resin under vacuum.

3. The method for three-dimensional reconstruction of porous Ni-YSZ electrodes for solid oxide batteries according to claim 1, characterized in that: Step S2 includes: using an argon ion polisher to polish the sample to a flat cross section and thin the surface electrolyte.

4. The method for three-dimensional reconstruction of porous Ni-YSZ electrodes for solid oxide batteries according to claim 1, characterized in that: Step S3 includes the following sub-steps: S301, plating a Pt film on the sample surface; S302, fixing the sample to a nail table with a pressing sheet; S303, depositing a Pt protective layer in the target area; S304: Processing grooves on both sides of the target area.

5. The method for three-dimensional reconstruction of porous Ni-YSZ electrodes for solid oxide batteries according to claim 4, characterized in that: Step S303 includes: sending the sample to a dual-beam electron microscope for processing and observation, using ion beam induced deposition technology to deposit a Pt protective layer on the surface of the target area, tilting the stage, adjusting the working distance to the concentric point, and using a Ga+ ion beam to perform cross-section processing. To ensure the quality of the SEM image, tilt compensation and dynamic focusing functions are used to correct the geometric distortion of the image.

6. The method for three-dimensional reconstruction of porous Ni-YSZ electrodes for solid oxide batteries according to claim 4, characterized in that: Step S304 includes: using a higher ion beam current to cut a large groove around the target area to avoid detector shadowing during SEM imaging, and then reducing the ion beam current to a set value to perform fine processing of the final cross section.

Images