TSV etching equipment fault identification method based on multi-dimensional parameter verification

Abstract

The present invention provides a TSV etching equipment fault identification method based on multi-dimensional parameter verification, which relates to the technical field of fault identification based on measuring electrical variables. In the process of operating the TSV etching equipment to perform etching on the etching object using the equipment operating parameters, an image acquisition node set is used to control the scanning electron microscope to perform image acquisition to obtain an etching dynamic image sequence; the etching dynamic image sequence is used to traverse and compare the twin etching finished products to locate the etching deviation, and based on the positioning result, the data recorder of the TSV etching equipment is reversely called to obtain the measured control parameter set; the etching deviation analysis is performed to generate the equipment fault identification result. The technical problem that the TSV etching equipment fault type cannot be diagnosed quickly and accurately in the prior art, resulting in a long time-consuming operation and maintenance of the TSV equipment fault, is solved. The technical effect of achieving rapid and accurate diagnosis and identification of TSV etching equipment faults and providing a reference for subsequent equipment operation and maintenance is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault identification based on measuring electrical variables, and in particular to a TSV etching equipment fault identification method based on multi-dimensional parameter verification. Background Art

[0002] In existing TSV (Through-Silicon Via) etching technology, equipment fault diagnosis often relies on manual experience and regular maintenance, which may result in fault diagnosis not being timely and accurate enough.

[0003] For example, when a TSV etching device fails during production, the operator may first check the physical components of the device, such as pumps and valves, and ignore potential problems with the control system. Due to the lack of fast and accurate diagnostic tools, this can lead to a time-consuming troubleshooting process, sometimes even requiring the disassembly and reassembly of the device, thus extending downtime.

[0004] In addition, if the failure is caused by software errors or improper parameter settings in the control system, and these problems are not discovered and resolved in a timely manner, it may lead to a decline in product quality, such as uneven etching depth or shape deviation, which in turn affects the product yield and market competitiveness.

[0005] In summary, the prior art has a technical problem that it is unable to quickly and accurately diagnose the type of TSV etching equipment failure, resulting in a long time-consuming operation and maintenance of TSV equipment failures, which indirectly affects production efficiency and product quality. Summary of the invention

[0006] The present invention provides a TSV etching equipment fault identification method based on multi-dimensional parameter verification, which is used to solve the technical problem in the prior art that the fault type of TSV etching equipment cannot be quickly and accurately diagnosed, resulting in a long time-consuming operation and maintenance of TSV equipment faults, which indirectly affects production efficiency and product quality.

[0007] In view of the above problems, the present invention provides a TSV etching equipment fault identification method based on multi-dimensional parameter verification, the method comprising: interactively obtaining equipment operating parameters of the TSV etching equipment, wherein the equipment operating parameters include etching paths and etching process parameters, wherein the TSV etching equipment is pre-configured with a scanning electron microscope; interactively obtaining a target silicon wafer design of a target etching object, and performing an etching process simulation according to the target silicon wafer design and the equipment operating parameters to obtain a twin etching product; interactively obtaining a field of view constraint of the scanning electron microscope, and obtaining image acquisition according to the twin etching product and the field of view constraint positioning. Node set; in the process of operating the TSV etching equipment with the equipment operating parameters to perform etching on the target etching object, the image acquisition node set is used to control the scanning electron microscope to perform image acquisition to obtain an etching dynamic image sequence; the etching dynamic image sequence is used to traverse and compare the twin etching finished product to locate the etching deviation, and based on the positioning result, the data recorder of the TSV etching equipment is reversely called to obtain the measured control parameter set; based on the measured control parameter set, etching deviation analysis is performed to generate equipment fault identification results, and fault operation and maintenance of the TSV etching equipment is performed according to the equipment fault identification results.

[0008] In a possible implementation, a target silicon wafer design of a target etching object is interactively obtained, and an etching process simulation is performed according to the target silicon wafer design and the equipment operating parameters to obtain a twin etching product, and the following processing is performed:

[0009] Call target size information, target shape information and target material information from the target silicon wafer design; pre-build a material property library, and use the target material information to traverse the material property library to obtain target material properties; perform twin modeling according to the target size information and target shape information to obtain a twin etching object; define the twin etching object using the target material properties, and use the equipment operating parameters to perform etching process simulation on the twin etching object to obtain the twin etching product.

[0010] In a possible implementation, the field of view constraint of the scanning electron microscope is interactively obtained, and an image acquisition node set is obtained according to the twin etching product and the field of view constraint positioning, and the following processing is performed:

[0011] Fitting the etching path to the twin etching product to obtain a twin space path; dividing the twin space path using the 1 / 4 mark of the field of view constraint to obtain a plurality of space image acquisition nodes; the plurality of space image acquisition nodes constitute the image acquisition node set.

[0012] In a possible implementation, the etching dynamic image sequence is used to traverse and compare the twin etching finished product to locate the etching deviation, and based on the positioning result, the data recorder of the TSV etching equipment is reversely called to obtain the measured control parameter set, and the following processing is performed:

[0013] A global etching twin image is acquired from the twin etching finished product according to the field of view constraint and the image acquisition node set; the etching dynamic image sequence is aligned and spliced ​​according to the global etching twin image to obtain a global etching dynamic image; an etching deviation recognition model is pre-constructed, and the etching deviation recognition model is used to traverse and compare the global etching twin image and the global etching dynamic image to locate the etching deviation node set; a data capture window is preset, and the data capture window is used as a constraint, and the measured control parameter set is obtained by calling the data recorder of the TSV etching equipment according to the etching deviation node set.

[0014] In a possible implementation, an etching deviation recognition model is pre-constructed, and the etching deviation recognition model is used to traverse and compare the global etching twin image and the global etching dynamic image to locate the etching deviation node set, and the following processing is performed:

[0015] The etching deviation recognition model includes an image alignment segmentation layer and an etching deviation recognition layer; the global etching twin image and the global etching dynamic image are synchronized to the image alignment segmentation layer to perform mapping segmentation to obtain multiple groups of etching verification images; the multiple groups of etching verification images are synchronized to the etching deviation recognition layer to perform etching deviation judgment to obtain K groups of etching verification images that do not meet the preset etching deviation threshold; K etching deviation nodes are reversely located in the twin etching finished product according to the K groups of etching verification images, and the K etching deviation nodes constitute the etching deviation node set.

[0016] In a possible implementation, etching deviation analysis is performed based on the measured control parameter set to generate a device fault identification result, and the following processing is performed:

[0017] The measured control parameter set is used to locally replace the etching process parameters in the equipment operating parameters to obtain updated operating parameters; the etching process verification is performed according to the target silicon wafer design and the updated operating parameters to obtain updated etching products; the etching deviation identification model is scheduled to perform deviation type analysis on the updated etching products and etching deviation node set to obtain the equipment fault identification result.

[0018] In a possible implementation, the etching deviation identification model is scheduled to perform deviation type analysis on the updated etching product and the etching deviation node set, obtain the equipment fault identification result, and perform the following processing:

[0019] According to the field of view constraint and the image acquisition node set, a global updated twin image is acquired in the updated etching finished product; the etching deviation recognition model is used to traverse and compare the global updated twin image and the global etching dynamic image; if the etching deviation node set output by the etching deviation recognition model is an empty set, the equipment fault recognition result is an equipment control fault. If the etching deviation node set output by the etching deviation recognition model is a non-empty set, the equipment fault recognition result is an equipment mechanical fault.

[0020] The technical solution provided in this application has at least the following technical effects or advantages:

[0021] The method provided in the embodiment of the present application obtains equipment operating parameters of a TSV etching device by interaction, wherein the equipment operating parameters include an etching path and etching process parameters, wherein the TSV etching device is pre-configured with a scanning electron microscope; interactively obtains a target silicon wafer design of a target etching object, and performs an etching process simulation according to the target silicon wafer design and the equipment operating parameters to obtain a twin etching product; interactively obtains the field of view constraint of the scanning electron microscope, and locates and obtains an image acquisition node set according to the twin etching product and the field of view constraint; in the process of operating the TSV etching device to perform etching on the target etching object using the equipment operating parameters, uses the image acquisition node set to control the scanning electron microscope to perform image acquisition to obtain an etching dynamic image sequence; uses the etching dynamic image sequence to traverse and compare the twin etching product to locate the etching deviation, and based on the positioning result, reversely calls the data recorder of the TSV etching device to obtain a measured control parameter set; performs etching deviation analysis based on the measured control parameter set to generate an equipment fault identification result, and performs fault operation and maintenance of the TSV etching device according to the equipment fault identification result. By collecting and verifying multi-dimensional parameters such as etching process parameters, etching product information, and etching twin modeling information, we can quickly and accurately diagnose and identify TSV etching equipment faults, provide a reference for subsequent equipment operation and maintenance, thereby improving production efficiency, reducing downtime, and ensuring the technical effect of product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic flow chart of a TSV etching equipment fault identification method based on multi-dimensional parameter verification provided by the present invention;

[0023] Figure 2 A schematic diagram of the process of obtaining twin etching products in the TSV etching equipment fault identification method based on multi-dimensional parameter verification provided by the present invention. DETAILED DESCRIPTION

[0024] The present invention provides a TSV etching equipment fault identification method based on multi-dimensional parameter verification, which is used to solve the technical problem that the TSV etching equipment fault type cannot be quickly and accurately diagnosed in the prior art, resulting in a long time-consuming operation and maintenance of TSV equipment faults, which indirectly affects production efficiency and product quality. By collecting and verifying multi-dimensional parameters of etching process parameters, etching finished product information, and etching twin modeling information, it is possible to quickly and accurately perform TSV etching equipment fault diagnosis and identification, provide a reference for subsequent equipment operation and maintenance, thereby improving production efficiency, reducing downtime, and ensuring the technical effect of product quality.

[0025] Below, the technical solutions in the present invention will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments described herein. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. It should also be noted that, for the convenience of description, only the parts related to the present invention are shown in the accompanying drawings, rather than all of them.

[0026] like Figure 1 As shown, the present invention provides a TSV etching equipment fault identification method based on multi-dimensional parameter verification, the method comprising:

[0027] A100: Interactively obtain equipment operating parameters of a TSV etching equipment, wherein the equipment operating parameters include an etching path and etching process parameters, wherein the TSV etching equipment is pre-configured with a scanning electron microscope.

[0028] Specifically, it should be understood that TSV (Through-Silicon Via) etching equipment is a precision equipment used in the semiconductor manufacturing process, which is specially used to make vertical interconnection channels on silicon wafers. TSV technology is mainly used in the manufacture of 3D integrated circuits, and TSV technology is used to achieve vertical interconnection between multi-layer silicon wafers, thereby improving the integration and performance of the circuit.

[0029] In this embodiment, by interacting with the control system of the TSV etching device, the etching path and etching process parameters currently input into the TSV etching device are collected to form the device operation parameters.

[0030] The etching path refers to the moving track of the etching tool on the silicon wafer during the etching process. The etching path is used to ensure that the etching process is carried out according to the predetermined design. The etching process parameters include etching rate, temperature, type and concentration of chemical agents, pressure, etc. These parameters affect the depth, uniformity and quality of etching.

[0031] It should be understood that among the above multi-index etching process parameters, some parameters are constant and some parameters are variable. For example, for a more complex etching process, some parameters may need to be adjusted according to the etching depth, silicon wafer material properties or etching time. As the etching process proceeds, the concentration of the chemical agent may need to be adjusted in time to maintain the optimal etching rate. For example, if a specific chemical agent concentration and temperature are proven to produce the best etching effect, they may remain constant throughout the etching process.

[0032] In this embodiment, the scanning electron microscope is a high-resolution imaging tool that can provide detailed images of the silicon wafer surface and capture images of the silicon wafer surface in real time during the etching process to monitor the etching quality and progress. By analyzing the images obtained by the scanning electron microscope, defects or deviations that may occur in the etching process can be detected, and the cause of the problem can be diagnosed based on the defects and deviations, such as equipment failure or improper process parameters.

[0033] A200: interactively obtain a target silicon wafer design of a target etching object, and perform an etching process simulation according to the target silicon wafer design and the equipment operating parameters to obtain a twin etching product.

[0034] In one embodiment, Figure 2 As shown, the target silicon wafer design of the target etching object is interactively obtained, and the etching process simulation is performed according to the target silicon wafer design and the equipment operating parameters to obtain a twin etching product. The method step A200 provided by the present invention also includes:

[0035] A210: Calling target size information, target shape information and target material information from the target silicon wafer design.

[0036] A220: Pre-build a material property library, and use the target material information to traverse the material property library to obtain the target material properties.

[0037] A230: Perform twin modeling according to the target size information and target shape information to obtain a twin etching object.

[0038] A240: Using the target material characteristics to define the twin etching object, and using the equipment operating parameters to perform etching process simulation on the twin etching object to obtain the twin etching product.

[0039] Specifically, in this embodiment, the target etching object is an unspecified silicon wafer that is currently planned to be processed using TSV etching equipment. By interacting with a design database or a design system, the target silicon wafer design of the target etching object is obtained. The target silicon wafer design is detailed design data of the silicon wafer, specifically including key information such as target size information (such as diameter, thickness, etc.), target shape information (such as round, rectangular, etc.) and target material information (such as silicon purity, doping type, etc.).

[0040] A material property library containing various material properties is pre-established to store the physical, chemical and mechanical properties of different materials. The target material information is used as an index to retrieve the corresponding material properties in the material property library to obtain the target material properties.

[0041] Based on conventional modeling methods, the target size information and target shape information are used to create the twin etching object in the CAD software to restore the three-dimensional geometric shape of the target etching object. Then, the twin etching object is imported into the FEA (finite element analysis) software, and material properties such as Young's modulus, Poisson's ratio, and thermal expansion coefficient are assigned to the twin etching object according to the target material properties.

[0042] After defining the material properties, the twin etching object is meshed and a finite element mesh is created to ensure that the key areas (such as the expected etching areas) have sufficient mesh density. Then, the etching physical process of the twin etching object is simulated according to the equipment operating parameters. Physical phenomena such as chemical reactions, thermal effects, fluid flow, etc. are considered in the simulation to improve the degree of restoration of the simulated etching results to the actual etching results.

[0043] FEA (finite element analysis) software gradually removes silicon wafer material of the twin etching object according to equipment operating parameters to generate a virtual representation of the twin etching product.

[0044] This embodiment achieves the technical effect of generating a virtual representation of a twin etching product that meets the design requirements based on finite element analysis, thereby providing favorable predictions and guidance for actual production.

[0045] A300: interactively obtain the field of view constraint of the scanning electron microscope, and obtain an image acquisition node set based on the twin etching product and the field of view constraint positioning.

[0046] In one embodiment, the field of view constraint of the scanning electron microscope is interactively obtained, and the image acquisition node set is obtained according to the twin etching product and the field of view constraint positioning. The method step A300 provided by the present invention also includes:

[0047] A310: Fit the etching path to the twin etching product to obtain a twin space path.

[0048] A320: Use the 1 / 4 mark of the field of view constraint to divide the twin space path to obtain multiple space image acquisition nodes.

[0049] A330: The multiple spatial image acquisition nodes constitute the image acquisition node set.

[0050] Specifically, in this embodiment, the field of view constraint of the scanning electron microscope is obtained interactively. The field of view constraint is the actual size of the area that can be observed by the scanning electron microscope. The field of view constraint is used to ensure that the collected image can cover the etching processing area. In this embodiment, the field of view constraint is a square.

[0051] Based on the field of view constraint, this embodiment uses numerical fitting technology to fit the actual etching path with the theoretically designed twin etching product to obtain the twin space path with spatial folding. The twin space path is used to ensure that the etching simulation path is consistent with the expected processing path.

[0052] Furthermore, 1 / 4 of the length or width of the field of view constraint is used as the path length division constraint. According to the path length constraint, the twin space path is equally spaced into multiple sections, and the length of each section is equal to 1 / 4 of the length or width of the field of view.

[0053] Then, the center of each divided section is used as an image acquisition node to obtain multiple spatial image acquisition nodes, which will be the key areas of scanning electron microscope imaging. The multiple spatial image acquisition nodes constitute the image acquisition node set. Based on the image acquisition node set, the target etching object image of the etching process is acquired, and an image covering all etching processing traces can be obtained.

[0054] This embodiment achieves the technical effect of ensuring that the collected image can cover the entire etching processing area while meeting the imaging requirements of the scanning electron microscope.

[0055] A400: In the process of operating the TSV etching equipment with the equipment operating parameters to perform etching on the target etching object, the image acquisition node set is used to control the scanning electron microscope to perform image acquisition to obtain an etching dynamic image sequence.

[0056] Specifically, in this embodiment, according to the equipment operating parameters (including etching path and etching process parameters) interactively obtained in the previous steps, the operating conditions of the TSV etching equipment are set, and the target etching object is placed in the TSV etching equipment to prepare for etching.

[0057] The twin space path and image acquisition node set are used as the image acquisition navigation of the etching process. During the etching process of the TSV etching equipment, the scanning electron microscope is controlled to sequentially acquire the etching process image of each node according to the navigation instructions of the image acquisition node set to obtain the etching dynamic image sequence.

[0058] A500: Use the etching dynamic image sequence to traverse and compare the twin etching finished product to locate the etching deviation, and based on the positioning result, reversely call the data recorder of the TSV etching equipment to obtain the measured control parameter set.

[0059] In one embodiment, the etching dynamic image sequence is used to traverse and compare the twin etching finished product to locate the etching deviation, and based on the positioning result, the data recorder of the TSV etching equipment is reversely called to obtain the measured control parameter set. The method step A500 provided by the present invention also includes:

[0060] A510: Acquire a global etching twin image from the twin etching finished product according to the field of view constraint and the image acquisition node set.

[0061] A520: Align and splice the etching dynamic image sequence according to the global etching twin image to obtain a global etching dynamic image.

[0062] A530: Pre-build an etching deviation recognition model, and use the etching deviation recognition model to traverse and compare the global etching twin image and the global etching dynamic image to locate the etching deviation node set.

[0063] A540: Preset a data capture window, and use the data capture window as a constraint to obtain the measured control parameter set by calling the data recorder of the TSV etching equipment according to the etching deviation node set.

[0064] In one embodiment, an etching deviation recognition model is pre-constructed, and the etching deviation recognition model is used to traverse and compare the global etching twin image and the global etching dynamic image to locate the etching deviation node set. Step A530 of the method provided by the present invention also includes:

[0065] A531: The etching deviation recognition model includes an image alignment segmentation layer and an etching deviation recognition layer.

[0066] A532: Synchronize the global etching twin image and the global etching dynamic image to the image alignment segmentation layer to perform mapping segmentation to obtain multiple sets of etching verification images.

[0067] A533: Synchronize the multiple groups of etching verification images to the etching deviation identification layer to perform etching deviation judgment, and obtain K groups of etching verification images that do not meet the preset etching deviation threshold.

[0068] A534: Reversely locate K etching deviation nodes in the twin etching product according to the K groups of etching verification images, and the K etching deviation nodes constitute the etching deviation node set.

[0069] Specifically, in this embodiment, according to the field of view constraints of the scanning electron microscope and the order relationship of multiple image acquisition nodes in the image acquisition node set, virtual image acquisition is performed on the twin model to obtain multiple etching twin images that are consistent with the acquisition positions of multiple etching dynamic images in the etching dynamic image sequence, and then the multiple etching twin images are adjacently overlapped and spliced ​​to obtain the global etching twin image. The global etching twin image covers the etching processing part of the entire twin etching product, providing a comprehensive view of the entire etching process.

[0070] Feature points, such as edges, corners or other significant visual features, are extracted from multiple etching dynamic images in the etching dynamic image sequence, and then feature matching technology, such as SIFT (scale-invariant feature transform) or ORB (Oriented FAST and Rotated BRIEF), is used to align the images in the image sequence to ensure their spatial consistency. According to the feature point matching results, the spatial transformation parameters between images, such as translation and rotation, are calculated. The spatial transformation parameters are applied to splice the multiple etching dynamic images to form a continuous global etching dynamic image. At the same time, the global etching dynamic image is compared with the twin model in image length to ensure that the global etching dynamic image completely covers the entire process of the physical etching process.

[0071] A pre-built etching deviation recognition model is used to realize automatic deviation detection. The etching deviation recognition model includes an image alignment segmentation layer and an etching deviation recognition layer. The specific use method of the etching deviation recognition model is as follows:

[0072] The image alignment and segmentation layer is responsible for receiving and aligning the global etching twin image and the global etching dynamic image, and mapping and segmenting the global etching twin image and the global etching dynamic image after image alignment to obtain multiple sets of etching verification images with corresponding relationships. The purpose of image segmentation is to reduce the amount of data for image detail comparison and reduce the computing power requirements for subsequent deviation analysis.

[0073] The etching deviation identification layer performs quantitative analysis of etching deviation. The specific quantitative analysis method is to interactively obtain multiple sample etching dynamic images, multiple sample etching twin images and multiple sample etching deviation coefficients, wherein each sample etching deviation coefficient is used to evaluate the similarity of a set of sample etching dynamic images - sample etching twin images, and each sample etching deviation coefficient is determined by etching shape analysis, and the data is relatively reliable.

[0074] The multiple sample etching dynamic images, multiple sample etching twin images and multiple sample etching deviation coefficients are used as training data, and the conventional training method for training back-propagation neural network is adopted to train the etching deviation recognition layer constructed by back-propagation grid, so that the etching deviation recognition layer can perform effective analysis and output of image similarity.

[0075] At the same time, the etching deviation recognition layer threshold is inserted with an etching deviation threshold, and the multiple groups of etching verification images are synchronized to the etching deviation recognition layer to perform etching deviation judgment, and K groups of etching verification images that do not meet the preset etching deviation threshold are obtained. It should be understood that the K groups of etching verification images indicate that there are etching deviations that affect the performance of the finished etching product. According to the K groups of etching verification images, K etching deviation nodes are reversely located in the twin etching product. The K etching deviation nodes are specific positions or areas that have deviations or do not meet the expected quality standards during the etching process, and the K etching deviation nodes have K etching time identifiers. The K etching deviation nodes constitute the etching deviation node set.

[0076] In this embodiment, a data capture window is preset, and the data capture window is a time interval. For example, the time interval of the data capture window is 1 / 40 of the total etching time of the equipment operating parameters in step A100. K measured control parameters are obtained by calling the data recorder of the TSV etching equipment according to the K etching time identifiers of the K etching deviation nodes and the data capture window to form the measured control parameter set.

[0077] The time spans of the K measured control parameters all satisfy the data capture window, and the beginnings of the K measured control parameters are the K etching time identifiers, and the data structure of each measured control parameter is a local etching process parameter.

[0078] This embodiment achieves the technical effect of locating the etching deviation node and the specific etching process parameters causing the etching deviation, and providing a reference for the subsequent verification and determination of the cause of the etching deviation.

[0079] A600: Perform etching deviation analysis based on the measured control parameter set, generate equipment fault identification results, and perform fault operation and maintenance of the TSV etching equipment according to the equipment fault identification results.

[0080] In one embodiment, etching deviation analysis is performed based on the measured control parameter set to generate equipment fault identification results. The method step A600 provided by the present invention further includes:

[0081] A610: Use the measured control parameter set to locally replace the etching process parameters in the equipment operating parameters to obtain updated operating parameters.

[0082] A620: Perform etching process verification according to the target silicon wafer design and the updated operating parameters to obtain an updated etching product.

[0083] A630: Scheduling the etching deviation identification model to perform deviation type analysis on the updated etching product and etching deviation node set to obtain the equipment fault identification result.

[0084] In one embodiment, the etching deviation identification model is scheduled to perform deviation type analysis on the updated etching product and the etching deviation node set to obtain the equipment fault identification result. Step A630 of the method provided by the present invention further includes:

[0085] A631: Obtain a globally updated twin image in the updated etching finished product according to the field of view constraints and the image acquisition node set.

[0086] A632: Use the etching deviation recognition model to traverse and compare the global update twin image and the global etching dynamic image.

[0087] A633: If the etching deviation node set output by the etching deviation identification model is an empty set, the equipment fault identification result is an equipment control fault.

[0088] In one embodiment, the method steps provided by the present invention further include:

[0089] A634: If the etching deviation node set output by the etching deviation identification model is a non-empty set, then the equipment fault identification result is an equipment mechanical failure.

[0090] Specifically, based on step A500, it can be known that the time spans of the K measured control parameters all meet the data capture window, and the start of the K measured control parameters is the K etching time identifiers, and the data structure of each measured control parameter is a local etching process parameter.

[0091] In this embodiment, K data replacement windows are located in the equipment operating parameters according to K etching time identifiers and data capture windows, and then K measured control parameters are used in the K data replacement window mappings to partially replace the etching process parameters in the original equipment operating parameters to obtain updated operating parameters, which reflect the specific process conditions that may cause deviations.

[0092] The same method as previously described for generating twin etching products is used to perform etching process verification according to the target silicon wafer design and the updated operating parameters to obtain updated etching products.

[0093] The same method of obtaining the global etching twin image is adopted, and the global updated twin image is obtained by collecting the updated etching finished product according to the field of view constraint and the image acquisition node set.

[0094] Referring to the same method as above, in which the etching deviation recognition model is used to traverse and compare the global etching twin image and the global etching dynamic image to locate the etching deviation node set, the etching deviation recognition model is used to traverse and compare the global update twin image and the global etching dynamic image.

[0095] If the etching deviation node set output by the etching deviation identification model is an empty set, it means that the current etching result is consistent with the etching control parameters actually recorded by the TSV etching equipment. The cause of the etching deviation is that the etching control of the TSV etching equipment is inconsistent with the pre-set etching process parameters, which points to a control failure. The equipment fault identification result is an equipment control failure.

[0096] On the contrary, if the etching deviation node set output by the etching deviation identification model is a non-empty set, it means that the deviation is related to the performance problem of the mechanical component, and the equipment fault identification result is an equipment mechanical fault.

[0097] This embodiment distinguishes the fault types by analyzing the etching deviation node set, thereby providing a reference for taking appropriate maintenance or adjustment measures.

[0098] This implementation achieves rapid and accurate TSV etching equipment fault diagnosis and identification through multi-dimensional parameter collection and verification of etching process parameters, etching finished product information, and etching twin modeling information, providing a reference for subsequent equipment operation and maintenance, thereby improving production efficiency, reducing downtime, and ensuring the technical effect of product quality.

[0099] Any of the methods or steps described above may be stored as computer instructions or programs in various types of computer memories, and the computer instructions or programs may be recognized by various types of computer processors to implement any of the methods or steps described above.

[0100] Based on the above specific embodiments of the present invention, any improvements and modifications made to the present invention by those skilled in the art without departing from the principles of the present invention shall fall within the patent protection scope of the present invention.

Claims

1. A TSV etching equipment fault identification method based on multi-dimensional parameter verification, characterized in that: The method comprises: Interactively obtaining equipment operating parameters of a TSV etching equipment, wherein the equipment operating parameters include an etching path and an etching process parameter, wherein the TSV etching equipment is pre-configured with a scanning electron microscope; Interactively obtain a target silicon wafer design of a target etching object, and perform an etching process simulation according to the target silicon wafer design and the equipment operating parameters to obtain a twin etching product; Interactively obtain the field of view constraint of the scanning electron microscope, and obtain an image acquisition node set according to the twin etching product and the field of view constraint positioning; In the process of operating the TSV etching device to perform etching on the target etching object using the device operating parameters, the image acquisition node set is used to control the scanning electron microscope to perform image acquisition to obtain an etching dynamic image sequence; The etching dynamic image sequence is used to traverse and compare the twin etching finished product positioning etching deviation, and based on the positioning result, the data recorder of the TSV etching equipment is reversely called to obtain the measured control parameter set; An etching deviation analysis is performed based on the measured control parameter set to generate an equipment fault identification result, and fault operation and maintenance of the TSV etching equipment is performed according to the equipment fault identification result.

2. The TSV etching equipment fault identification method based on multi-dimensional parameter verification according to claim 1, characterized in that: Interactively obtaining a target silicon wafer design of a target etching object, and performing an etching process simulation according to the target silicon wafer design and the equipment operating parameters to obtain a twin etching product, the method further comprising: Retrieving target size information, target shape information and target material information from the target silicon wafer design; Pre-constructing a material property library, and using the target material information to traverse the material property library to obtain the target material properties; Performing twin modeling according to the target size information and the target shape information to obtain a twin etching object; The target material properties are used to define the twin etching object, and the equipment operating parameters are used to perform an etching process simulation on the twin etching object to obtain the twin etching product.

3. The TSV etching equipment fault identification method based on multi-dimensional parameter verification according to claim 2, characterized in that: Interactively obtaining the field of view constraint of the scanning electron microscope, and obtaining an image acquisition node set according to the twin etching product and the field of view constraint positioning, the method further comprising: Fitting the etching path to the twin etching product to obtain a twin space path; Dividing the twin space path by using a 1 / 4 mark of the field of view constraint to obtain a plurality of space image acquisition nodes; The multiple space image acquisition nodes constitute the image acquisition node set.

4. The TSV etching equipment fault identification method based on multi-dimensional parameter verification according to claim 1, characterized in that: The etching dynamic image sequence is used to traverse and compare the twin etching finished product positioning etching deviation, and based on the positioning result, the data recorder of the TSV etching equipment is reversely called to obtain the measured control parameter set, and the method further includes: Acquire a global etching twin image from the twin etching finished product according to the field of view constraint and the image acquisition node set; Aligning and splicing the etching dynamic image sequence according to the global etching twin image to obtain a global etching dynamic image; Pre-constructing an etching deviation recognition model, and using the etching deviation recognition model to traverse and compare the global etching twin image and the global etching dynamic image to locate the etching deviation node set; A data capture window is preset, and the data capture window is used as a constraint, and the measured control parameter set is obtained by calling the data recorder of the TSV etching device according to the etching deviation node set.

5. The TSV etching equipment fault identification method based on multi-dimensional parameter verification according to claim 4, characterized in that: Pre-constructing an etching deviation recognition model, and using the etching deviation recognition model to traverse and compare the global etching twin image and the global etching dynamic image to locate the etching deviation node set, the method further includes: The etching deviation recognition model includes an image alignment segmentation layer and an etching deviation recognition layer; Synchronize the global etching twin image and the global etching dynamic image to the image alignment segmentation layer to perform mapping segmentation to obtain multiple groups of etching verification images; Synchronizing the plurality of etching verification images to the etching deviation identification layer to perform etching deviation judgment, and obtaining K groups of etching verification images that do not meet a preset etching deviation threshold; K etching deviation nodes are reversely located on the twin etching product according to the K groups of etching verification images, and the K etching deviation nodes constitute the etching deviation node set.

6. The TSV etching equipment fault identification method based on multi-dimensional parameter verification according to claim 5, characterized in that: Performing etching deviation analysis based on the measured control parameter set to generate equipment fault identification results, the method further includes: locally replacing the etching process parameters in the equipment operating parameters with the measured control parameter set to obtain updated operating parameters; Performing etching process verification according to the target silicon wafer design and the updated operating parameters to obtain an updated etching product; The etching deviation identification model is scheduled to perform deviation type analysis on the updated etching product and the etching deviation node set to obtain the equipment fault identification result.

7. The TSV etching equipment fault identification method based on multi-dimensional parameter verification according to claim 6, characterized in that: Scheduling the etching deviation identification model to perform deviation type analysis on the updated etching product and the etching deviation node set to obtain the equipment fault identification result, the method further comprising: Acquire a globally updated twin image in the updated etching finished product according to the field of view constraint and the image acquisition node set; The etching deviation recognition model is used to traverse and compare the global updated twin image and the global etching dynamic image; If the etching deviation node set output by the etching deviation identification model is an empty set, the equipment fault identification result is an equipment control fault.

8. The TSV etching equipment fault identification method based on multi-dimensional parameter verification according to claim 7, characterized in that: If the etching deviation node set output by the etching deviation identification model is a non-empty set, the equipment fault identification result is an equipment mechanical fault.

Images