A method and system for detecting and analyzing heavy metal residues on a TSV wafer after CMP cleaning

Abstract

This application provides a method and system for detecting and analyzing heavy metal residues in TSV wafers after CMP cleaning, relating to the field of integrated circuit technology, and aims to improve the technical problem of inaccurate evaluation results of detection and analysis methods. The method includes: acquiring surface morphology data and surface heavy metal concentration distribution data of the wafer to be analyzed; dividing the surface of the wafer to be analyzed into multiple grid cells of a preset size based on the surface morphology data, and determining a terrain influence factor for each grid cell; generating an initial propagation contamination value for each grid cell based on the surface heavy metal concentration distribution data of the grid cell and its neighboring grid cells and the terrain influence factor of the neighboring grid cells; generating an initial contamination distribution map of the wafer to be analyzed based on the initial propagation contamination values ​​of all grid cells; and extracting a global contamination index characterizing the overall contamination level of the wafer from the initial contamination distribution map.

Description

Technical Field

[0001] This application relates to the field of integrated circuit technology, and in particular to a method and system for detecting and analyzing heavy metal residues in TSV wafers after CMP cleaning. Background Technology

[0002] In 3D integrated circuit (3D-IC) technology, through-silicon vias (TSVs) are a key structure for enabling vertical chip stacking and interconnection. The fabrication process of TSVs involves a series of complex processes, including deep silicon etching, deposition of insulating and barrier layers, and metal (typically copper) filling. After metal filling, chemical mechanical polishing (CMP) is widely used to remove excess metal from the wafer surface and achieve global planarization.

[0003] The CMP process itself introduces contaminants. CMP slurry typically contains chemical reagents, abrasive particles, and removed metal ions. Therefore, efficient post-CMP cleaning is essential to remove residual particles, organic matter, and metal contaminants from the wafer surface. Heavy metal residues (such as copper, iron, and tungsten) are a particularly serious problem. Even at extremely low concentrations, these metal ions can diffuse into active device regions, leading to gate oxide breakdown, increased leakage current, and device threshold voltage drift, severely impacting chip performance, reliability, and final yield.

[0004] Existing heavy metal residue detection technologies, such as total reflectance X-ray fluorescence spectroscopy (TXRF) or gas phase decomposition-inductively coupled plasma mass spectrometry (VPD-ICP-MS), while providing highly sensitive wafer-level average contamination concentrations, typically lack spatial resolution and cannot reveal the microscopic distribution characteristics of contaminants on the wafer surface. Imaging techniques, such as time-of-flight secondary ion mass spectrometry (ToF-SIMS), while providing high spatial resolution elemental distribution maps, typically offer qualitative or semi-quantitative results, and data interpretation heavily relies on operator experience.

[0005] More importantly, current technologies often treat heavy metal residue data as isolated chemical information when analyzing it, failing to effectively correlate it with the microscopic physical morphology of the wafer surface and key structural features such as TSVs. In fact, the microscopic morphology of the wafer surface after CMP (such as depressions, scratches, and surface roughness at TSV openings) significantly affects the flow behavior of the cleaning solution, creating fluid stagnation zones that lead to contaminant accumulation at these specific locations. Furthermore, the material and electrochemical properties of the TSV structure itself may also make it a preferential adsorption site for heavy metal ions. Summary of the Invention

[0006] This application provides a method and system for detecting and analyzing heavy metal residues in TSV wafers after CMP cleaning, which improves the technical problem that the detection and analysis methods in related technologies fail to effectively combine the microscopic physical characteristics of the wafer for quantitative evaluation, resulting in inaccurate evaluation results.

[0007] To achieve the above objectives, the embodiments of this application adopt the following technical solutions:

[0008] In a first aspect, this application provides a method for detecting and analyzing heavy metal residues in TSV wafers after CMP cleaning. The method includes: acquiring surface morphology data and surface heavy metal concentration distribution data of the wafer to be analyzed; dividing the surface of the wafer to be analyzed into multiple grid cells of a preset size based on the surface morphology data, and determining a terrain influence factor for each grid cell; for each grid cell, generating an initial propagation contamination value based on the surface heavy metal concentration distribution data of the grid cell and its neighboring grid cells and the terrain influence factor of the neighboring grid cells; generating an initial contamination distribution map of the wafer to be analyzed based on the initial propagation contamination values ​​of all grid cells; and extracting a global contamination index characterizing the overall contamination level of the wafer from the initial contamination distribution map.

[0009] In one possible implementation of the first aspect, the step of determining a terrain influence factor for each grid cell includes: for each grid cell, obtaining a fluid accumulation and a surface slope value based on the surface topography data; and obtaining the terrain influence factor of the grid cell based on the fluid accumulation and the surface slope value of the grid cell.

[0010] In one possible implementation of the first aspect, the terrain influence factor is obtained by the following formula:

[0011]

[0012] in, Let be the terrain influence factor of the grid cell in the i-th row and j-th column. The fluid accumulation amount of this grid cell, The surface slope value of the grid cell is given by k, m, and n, which are process coefficients associated with the chemical mechanical polishing process parameters.

[0013] In one possible implementation of the first aspect, the step of generating an initial propagation pollution value includes: constructing a weight kernel, wherein the value of each element in the weight kernel is determined by the terrain influence factor of the neighboring grid cell; and performing a convolution operation on the surface heavy metal concentration distribution data of the grid cell and its neighboring grid cells using the weight kernel to obtain the initial propagation pollution value.

[0014] In one possible implementation of the first aspect, the step of extracting a global contamination index characterizing the overall contamination level of the wafer from the initial contamination distribution map includes: statistically analyzing the numerical distribution of all initial propagated contamination values ​​in the initial contamination distribution map; and obtaining the values ​​at preset quantiles of the numerical distribution as the global contamination index.

[0015] In one possible implementation of the first aspect, the method further includes: after generating the initial contamination distribution map, modifying the initial contamination distribution map to obtain a modified contamination distribution map; wherein the step of extracting a global contamination index characterizing the overall contamination level of the wafer from the initial contamination distribution map specifically involves: extracting the global contamination index from the modified contamination distribution map.

[0016] In one possible implementation of the first aspect, the step of correcting the initial contamination distribution map includes: obtaining structural layout information of through-silicon vias (TSVs) on the wafer to be analyzed; determining a TSV proximity factor for each grid cell based on the structural layout information; generating a corrected contamination value based on the initial propagated contamination value and the TSV proximity factor for each grid cell, wherein the corrected contamination values ​​of all grid cells constitute the corrected contamination distribution map.

[0017] In one possible implementation of the first aspect, the via proximity factor characterizes the distance relationship between the grid cell and the nearest via edge.

[0018] Secondly, this application also provides a heavy metal residue detection and analysis system for TSV wafers after CMP cleaning, comprising: a data acquisition module for acquiring surface morphology data and surface heavy metal concentration distribution data of the wafer to be analyzed; an influence factor determination module for dividing the surface of the wafer to be analyzed into multiple grid cells of a preset size based on the surface morphology data, and determining a topographic influence factor for each grid cell; an initial value generation module for generating an initial propagation contamination value for each grid cell based on the surface heavy metal concentration distribution data of the grid cell and its neighboring grid cells and the topographic influence factor of the neighboring grid cells; a distribution map generation module for generating an initial contamination distribution map of the wafer to be analyzed based on the initial propagation contamination values ​​of all grid cells; and an index extraction module for extracting a global contamination index characterizing the overall contamination level of the wafer from the initial contamination distribution map.

[0019] In one possible implementation of the second aspect, it further includes: a correction module, used to correct the initial pollution distribution map after the distribution map generation module generates the initial pollution distribution map, to obtain a corrected pollution distribution map; wherein, the indicator extraction module is specifically used to: extract the global pollution indicator from the corrected pollution distribution map. Attached Figure Description

[0020] Figure 1 A flowchart illustrating a method for detecting and analyzing heavy metal residues in TSV wafers after CMP cleaning, provided for some embodiments of this application;

[0021] Figure 2 A schematic diagram of the structure of a heavy metal residue detection and analysis system for TSV wafers after CMP cleaning, provided for some embodiments of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0023] In the following description, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0024] Furthermore, in this application, directional terms such as "upper," "lower," "left," and "right" may be defined relative to the orientation of the components shown in the accompanying drawings. It should be understood that these directional terms can be relative concepts, used for relative description and clarification, and may change accordingly depending on the orientation of the components in the accompanying drawings.

[0025] In this application, unless otherwise expressly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral part; it can be a direct connection or an indirect connection through an intermediate medium. Furthermore, the term "electrical connection" can refer to the manner in which an electrical connection is used to achieve signal transmission.

[0026] As used herein, “about,” “approximately,” or “approximately” includes the stated value and a reference value within an acceptable range of deviation from the given value, characterized in that the acceptable range of deviation is determined by a person skilled in the art taking into account the measurement under discussion and the error associated with the measurement of the given quantity (i.e., the limitations of the measurement method).

[0027] This application provides a method for detecting and analyzing heavy metal residues on TSV wafers after CMP cleaning. This method aims to address the technical problems in existing technologies for detecting and assessing heavy metal residues on TSV wafer surfaces, including insufficient spatial resolution, failure to fully integrate the analysis of wafer surface microscopic physical characteristics, and difficulty in accurately reflecting the potential migration and accumulation risks of contaminants.

[0028] like Figure 1 As shown, the method includes:

[0029] S100: Obtain surface morphology data and surface heavy metal concentration distribution data of the wafer to be analyzed.

[0030] Topography data describes the three-dimensional microstructure of the wafer surface after CMP and cleaning. These microstructures, such as dishing around TSV openings, erosion of the protective layer material, scratches, and other surface irregularities, are key physical factors affecting the flow of CMP cleaning fluid and the residue of contaminant particles / ions. Those skilled in the art can use various high-resolution surface measurement techniques to obtain this data.

[0031] For example, an atomic force microscope (AFM), a scanning white light interferometer (SWLI), or an optical profilometer can be used to scan the entire or a specific area of ​​the wafer surface. The scan result is a series of three-dimensional coordinate points (x, y, z), where z represents the height value at position (x, y). This raw point cloud data will then be processed into a regular digital elevation model (DEM), i.e., a two-dimensional matrix where each element represents the height at the corresponding position.

[0032] Surface heavy metal concentration distribution data aims to characterize the spatial distribution information of specific heavy metal elements (such as Cu, Fe, W, etc.) on the wafer surface. Unlike traditional single-point or area averaging measurement methods, this application obtains a two-dimensional distribution map that spatially corresponds to surface morphology data.

[0033] For example, imaging-enabled surface elemental analysis techniques such as micro-X-ray fluorescence (μ-XRF) or time-of-flight secondary ion mass spectrometry (ToF-SIMS) can be employed. These techniques can scan the wafer surface at the micrometer scale and provide characteristic signal intensity or atomic concentration of the target element for each pixel. The analysis results are ultimately processed into a two-dimensional concentration matrix, where the value of each element represents the heavy metal concentration at the corresponding location. .

[0034] After data acquisition, data preprocessing is required, including noise filtering, coordinate alignment, and normalization, to ensure that the surface morphology data map and the heavy metal concentration distribution data map are precisely aligned in the spatial coordinate system, that is, any coordinate point (x,y) of the two data maps corresponds to the same physical location on the wafer.

[0035] S200. Based on the surface topography data, the surface of the wafer to be analyzed is divided into multiple grid cells of a preset size, and a terrain influence factor is determined for each grid cell.

[0036] The area covered by the preprocessed surface topography data (DEM) is divided into a region consisting of... A regular grid system composed of grid cells Each grid cell (in For row index, ; For column indexes, ) have the same size, for example, The grid size needs to be weighed based on the resolution of data acquisition and the required precision of analysis. After partitioning, surface morphology data and surface heavy metal concentration distribution data are mapped onto this grid system, with each grid cell... Corresponding to an average height value and an average measured concentration value .

[0037] A Topographical Influence Factor (TIF) is determined for each grid cell. The TIF is constructed based on the Stream Power Model, which describes the erosive power of water flow. It describes the potential for contaminant "retention or enrichment" due to local topographic variations as CMP cleaning fluid flows across the wafer's micro-surface. The more the topography of a region allows fluid to converge and the slower the flow velocity (e.g., in depressions), the higher the potential for contaminant enrichment, and the larger the TIF.

[0038] The specific steps for determining topographic influence factors are as follows:

[0039] S210. For each grid cell, based on the surface topography data, obtain a fluid accumulation amount and a surface slope value.

[0040] Flow accumulation is denoted as For grid cells In terms of its physical meaning, it means that when there is uniform "precipitation" (i.e., cleaning fluid) on the entire wafer surface, the "runoff" of all upstream grid cells converges at... The total amount. The higher this value, the more it indicates. It lies on a path where fluid converges or in a local depression. The fluid accumulation can be calculated by applying a flow direction algorithm (such as the D8 algorithm) to a digital elevation model (DEM). The D8 algorithm determines the unique, steepest neighboring cell for each cell as the flow direction, and then recursively accumulates the number of cells flowing through each cell from high to low, thus obtaining the fluid accumulation for each cell.

[0041] Surface slope value, denoted as , representing a grid cell The degree of surface inclination. Slope is a key factor affecting fluid velocity and energy. A flat area (low slope) is more likely to cause particulate matter settling and liquid residue than a steep area (high slope). Surface slope values ​​can be calculated using a numerical height model. exist The gradient at that point is used to obtain the gradient.

[0042] For example, the second-order finite difference method can be used to calculate:

[0043]

[0044]

[0045]

[0046] in, and These are the rates of change of height in the x and y directions, respectively. and It is the size of the grid cell. The slope value is expressed in radians.

[0047] S220. Based on the fluid accumulation and surface slope value of the grid cell, obtain the terrain influence factor of the grid cell.

[0048] Topographical Influence Factors Defined as a function combining fluid accumulation and surface slope, its form is borrowed from the flow power erosion model. This application defines it as follows:

[0049]

[0050] in, It is the first Line number Column grid cells The topographic influence factor is a dimensionless relative value that characterizes the pollution enrichment potential of the unit due to topographic factors. It is a unit The amount of fluid accumulation. It is a unit The surface slope value.

[0051] , , These are process coefficients. These coefficients are not variable parameters used for model fitting, but rather constants predetermined based on specific CMP process conditions. Their introduction tightly couples the model to the actual physical process. For example:

[0052] coefficient It is a basic proportionality coefficient that can be related to the viscosity, chemical composition, and other characteristics of the cleaning fluid.

[0053] index This reflects the degree of influence of fluid aggregation on contaminant enrichment. For slurries with higher viscosity or containing larger abrasive particles, the aggregation effect is more pronounced. The value may be larger. Typically... .

[0054] index This reflects the degree to which surface slope affects contaminant accumulation. Generally, during CMP cleaning, areas with gentler slopes (flatter surfaces) are more conducive to residue buildup. It could be a negative value, for example This indicates that the smaller the slope, the greater the influence factor. Alternatively, the formula can be... Replace with or ,at this time It is a positive value. To ensure the universality and monotonicity of the formula, it is defined here. For example, a metric representing "flatness" Then the formula can be written as ,at this time .

[0055] For example, for a specific Cu-CMP post-cleaning process, the cleaning solution used has a high viscosity and the wafer rotation speed is low. A set of definite process coefficients can be obtained through experimental calibration or theoretical derivation, for example... , , Once the coefficients are determined, this set of coefficients will be applied to the analysis of all wafers using the same process, ensuring the repeatability and determinism of the method.

[0056] S300: For each grid cell, based on the surface heavy metal concentration distribution data of the grid cell and its neighboring grid cells, and the topographic influence factor of the neighboring grid cells, an initial propagation pollution value is generated. S300 includes:

[0057] S310. Construct a weight kernel, wherein the value of each element in the weight kernel is determined by the terrain influence factor of the neighboring grid cell.

[0058] First, define a neighborhood range, for example, a The window for the target cell to be calculated. Centered on [the target location], a weight kernel of the same size is then constructed. The purpose of this kernel is to assign different weights to the measured concentrations of different cells within the neighborhood in subsequent calculations. The magnitude of the weight depends on the topographic influence factor of that neighborhood cell. This refers to its potential for pollution accumulation.

[0059] For example, for the target unit of Neighborhood (including itself), weight kernel Each element in (in (Relative coordinates within the neighborhood) can be defined as:

[0060]

[0061] in, The coordinates within the neighborhood are The formula essentially calculates the terrain influence factor of all units within the neighborhood. The values ​​are normalized to obtain a set of weights that sum to 1. A neighborhood unit's... The larger the value, the greater its weight in calculating the pollution value of the target unit, meaning that its pollution status has a greater impact on the target unit.

[0062] S320. Using the weight kernel, perform a convolution operation on the surface heavy metal concentration distribution data of the grid cell and its neighboring grid cells to obtain the initial propagation pollution value.

[0063] Obtain weight kernel Then, by comparing it with the surface heavy metal concentration measurement matrix of the corresponding neighborhood... Perform convolution to compute the target unit. The initial propagated contamination value is denoted as .

[0064]

[0065] In this formula, the unit The initial transmission contamination value is its The concentration is calculated as a weighted average of all measured concentrations within a neighborhood. The weights are dynamically determined by the topographic influence factor within that neighborhood. This process is performed on every grid cell across the entire wafer, starting from the original measured concentration matrix. A new matrix was generated that takes into account the pollution propagation effect under the influence of terrain. .

[0066] This allows for the smoothing of isolated high-concentration points caused by measurement noise, while simultaneously "enhancing" contamination signals located in high-retention-risk areas (high TIF). For example, a cell with a low measured concentration, if surrounded by multiple high-concentration, high-TIF neighbors, will have its calculated concentration reduced. The value will also increase accordingly, which more accurately reflects the harsh microenvironment in which it is located. Conversely, the influence of an isolated high-concentration measurement point located in a low-TIF region (which may be due to measurement error or irrelevant particles) will be "diluted" in the calculation by its surrounding low-TIF neighbors.

[0067] S400. Based on the initial propagation contamination values ​​of all grid cells, generate an initial contamination distribution map of the wafer to be analyzed.

[0068] After processing by S300, a mesh system similar to the original mesh system was obtained. Matrix of the same size Each element in this matrix This represents the corresponding grid cell. Pollution levels after taking into account topographically mediated local pollution propagation effects.

[0069] The process of generating the initial contamination distribution map is essentially the process of transforming this two-dimensional data matrix... Visualize it.

[0070] For example, a pseudo-color plot can be used to represent this. The numerical range in the matrix is ​​mapped onto a color map, for example, from blue (low contamination) to green, yellow, and finally red (high contamination). The resulting image can visually show the "hot spots" of heavy metal contamination on the wafer surface. These hot spots are not only areas with high concentrations in the original measurements, but also areas identified as having a high risk of contamination after being weighted by topographic factors.

[0071] This initial contamination distribution map is a key intermediate product of this method, providing deeper, more physically related information compared to the original concentration distribution map. For example, analysts can observe whether contamination is distributed along specific micro-gullies (corresponding to paths with high fluid accumulation) or concentrated in flat, concave areas around the TSV opening (corresponding to areas with high fluid accumulation and low slope).

[0072] S500. Extract a global contamination index from the initial contamination distribution map to characterize the overall contamination level of the wafer.

[0073] While contamination maps provide detailed spatial information, industrial production often requires one or more concise, quantifiable indicators to quickly determine whether the cleaning quality of an entire wafer is up to standard. This step aims to extract such a representative global contamination index from the complex distribution map.

[0074] The extraction of this indicator should avoid simple average calculations, as averages are easily lowered by a large number of low-pollution areas, thus masking the risk of localized severe pollution. This application proposes using an indicator based on statistical distribution.

[0075] S510: Statistically analyze the numerical distribution of all initial propagation pollution values ​​in the initial pollution distribution map.

[0076] This step is for All in the matrix Statistical analysis can be performed on these values, such as drawing histograms of these values, to understand the overall distribution of pollution levels, such as whether it is normally distributed, log-normally distributed, or has a significant heavy tail, the latter usually indicating the existence of a few extremely severe pollution points.

[0077] S520. Obtain the values ​​at preset quantiles of the numerical distribution as the global contamination index. Based on the statistical distribution in the previous step, select one or more high quantile values ​​as the global contamination index.

[0078] For example, the 95th percentile can be selected, denoted as . The physical meaning is that 95% of the area on the wafer surface has a contamination value below this value. Compared to the average or maximum value, It is not sensitive to extreme outliers (which may be noise), but it can effectively reflect the high level of contamination that is prevalent on wafers, making it a more robust and representative indicator for measuring yield and reliability.

[0079] Apart from Other indicators can also be extracted as needed, such as the 99th percentile. The proportion of grid cells with pollution values ​​exceeding a certain preset threshold. In this application, since this indicator is derived from a modified version... Extracted from data, rather than directly from raw measurement data. Extracted from [source], it is therefore more representative of the actual pollution risk.

[0080] In more sophisticated analytical scenarios, in addition to surface microstructure, specific artificial structures on the wafer, especially the TSV itself, also significantly influence the adsorption and residue of heavy metals. The sidewall material, filler metal, and complex geometry of the TSV openings can all become preferential adsorption sites for heavy metal ions. To take this factor into account, the method of this application further includes:

[0081] S600: After generating the initial pollution distribution map, the initial pollution distribution map is corrected to obtain a corrected pollution distribution map. S600 utilizes known TSV structure layout information to modify the initial pollution distribution map calculated based on terrain. A second correction is performed. For example, S600 includes:

[0082] S610. Obtain the structural layout information of the through-silicon vias on the wafer to be analyzed.

[0083] The structural layout information of TSVs typically comes from the chip's design layout files, such as GDSII or OASIS format files. From these files, the center coordinates, diameter, and possible array arrangements of each TSV can be precisely extracted. This information is then converted into an analysis mesh system. Aligned TSV position diagram.

[0084] S620. Based on the structural layout information, determine a silicon via proximity factor for each grid cell.

[0085] Through-Silicon Via Proximity Factor (TSV, TPF) is denoted as It is for each grid cell This is another quantitative parameter created. It is designed to characterize the additional contamination risk that the unit may have due to its proximity to the TSV structure. This risk may stem from the electrochemical potential difference between the TSV filler metal (such as Cu) and residual copper ions in the cleaning solution, or the chemical affinity of the TSV opening edge material for a particular metal.

[0086] The definition of TPF can take many forms, but its core is related to the distance to TSV.

[0087] For example, one way to define it is based on the distance from the center of the grid cell to the nearest TSV edge. The closer the distance, the greater the impact. This impact can be modeled using an exponential decay function:

[0088]

[0089] in, It is an attenuation constant whose value is related to the specific heavy metal species and the interaction strength of the materials used in the TSV (such as the Ta / TaN barrier layer, Cu seed layer, etc.). This value can be obtained through prior experimental data and is fixed for a specific process. For example, for Cu contamination, A larger value indicates a smaller but stronger influence range; for ions with stronger diffusion capabilities, The value may be small.

[0090] Another way to define it is to consider a certain radius. TSV density within the grid cell. Calculate the value with center and radius as... The proportion of the area occupied by TSV within the circular region is used as... This method is more suitable for areas with densely packed TSVs.

[0091] S630. Based on the initial propagation contamination value and the via proximity factor for each grid cell, a corrected contamination value is generated, and the corrected contamination values ​​of all grid cells constitute the corrected contamination distribution map.

[0092] Initial propagation contamination value and silicon via proximity factor The values ​​are combined to generate the final Corrected Contamination Value, denoted as... The combination should reflect the "risk enhancement" effect represented by TPF.

[0093] For example:

[0094]

[0095] in: It is the initial propagation contamination value obtained in S300. It is the silicon via proximity factor obtained in S620, and its value range is usually between [0,1]. It is a dimensionless weighting coefficient representing the contribution of the TSV structure itself to pollutant enrichment. This coefficient is also determined by the process and material properties and is a preset constant. For example, if it is known that the TSV material has extremely strong adsorption for the target heavy metal, You can choose a larger value, such as 0.5 or 1.0.

[0096] With this correction, the contamination value of grid cells close to the TSV will be reduced from the original value. It is magnified based on the original value, and the magnification ratio depends on its distance from the TSV (by...). The decision) and the adsorption risk of TSV itself (by (Decision). After performing this operation on all grid cells, the final corrected contamination distribution map is obtained, derived from the matrix. express.

[0097] S700. Extract the global pollution index from the modified pollution distribution map.

[0098] For example, calculation This ultimate global contamination indicator, because it incorporates both the hydrodynamic effects caused by microstructure (via TIF) and the electrochemical / chemical effects caused by TSV structure (via TPF), allows for a more comprehensive and accurate assessment of wafer cleaning quality.

[0099] This application also provides a heavy metal residue detection and analysis system, which is configured to perform the methods described in any of the above embodiments. The system can be a dedicated computer device or a semiconductor detection device integrated with analysis software. The system can be internally divided into multiple functional modules.

[0100] like Figure 2 As shown, the system includes:

[0101] The data acquisition module is used to interface with external measurement equipment (such as AFM, μ-XRF) to receive and acquire raw surface morphology data and surface heavy metal concentration distribution data of the wafer to be analyzed. It also performs preliminary data processing, such as format conversion, noise filtering, and coordinate alignment.

[0102] The influence factor determination module receives aligned surface topography data. Internally, it includes algorithms for mesh generation, as well as functions for calculating fluid accumulation (e.g., the D8 algorithm) and surface slope. Its core function is to determine the influence factor based on preset process coefficients (k,m,n) using formulas. Calculate the terrain influence factor for each grid cell .

[0103] The initial value generation module is used to receive the measured concentration matrix. and topographic influence factor matrix It is based on a preset neighborhood size (e.g., ), dynamically constructing weight kernels for each target grid and perform convolution operations. This generates the initial propagation contamination value matrix. .

[0104] The distribution map generation module is used to receive the initial propagation contamination value matrix. (or subsequent corrected contamination value matrix) It then calls the graphics rendering engine to convert the data into a pseudo-color pollution distribution map for users to view and analyze.

[0105] The indicator extraction module is used to process the pollution value matrix ( or Statistical analysis is performed to calculate the probability distribution of the values, and global pollution indicators are extracted based on preset parameters (such as 95% quantile). .

[0106] Correction Module (Optional): This module is also included in systems that implement a correction mechanism. It receives the design layout file (such as GDSII) and extracts the TSV location information. It adjusts the TSV position based on a preset attenuation constant. Calculate the proximity factor of through-silicon vias And according to the weighting coefficients Through formula The initial contamination values ​​are corrected, and the final corrected contamination value matrix is ​​output. Provide the distribution map generation module and the indicator extraction module.

[0107] These modules can be software program code segments, executed by one or more processors in the system. The processors can be configured to execute instructions stored in memory to implement the method steps described above. The memory can be any type of computer-readable medium.

[0108] The technical solution of this application, by creatively introducing and defining the Terrain Influence Factor (TIF) and Through-Silicon Via Proximity Factor (TPF) and integrating them into a deterministic analytical framework, achieves in-depth analysis of heavy metal residues on the surface of TSV wafers. This method not only identifies high-concentration contamination points but also reveals the intrinsic relationship between contamination distribution and the wafer's microscopic physical structure, thus providing more accurate and valuable guidance for optimizing post-CMP cleaning processes. Its output global contamination index, by integrating multiple key physical factors, can more reliably predict the final impact of contaminants on device performance and yield.

[0109] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0110] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0111] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another apparatus, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0112] The units described as separate components may or may not be physically separate. A component shown as a unit can be one physical unit or multiple physical units; that is, it can be located in one place or distributed in multiple different places. Depending on actual needs, some or all of the units can be selected to achieve the purpose of this embodiment.

[0113] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit described above can be implemented in hardware.

[0114] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for detecting and analyzing heavy metal residues in TSV wafers after CMP cleaning, characterized in that, The method includes: Acquire surface morphology data and surface heavy metal concentration distribution data of the wafer to be analyzed; Based on the surface topography data, the surface of the wafer to be analyzed is divided into multiple grid cells of a preset size, and a topography influence factor is determined for each grid cell; For each grid cell, an initial propagation pollution value is generated based on the surface heavy metal concentration distribution data of the grid cell and its neighboring grid cells and the topographic influence factor of the neighboring grid cells; Based on the initial propagation contamination values ​​of all grid cells, an initial contamination distribution map of the wafer to be analyzed is generated; Extract a global contamination index that characterizes the overall contamination level of the wafer from the initial contamination distribution map.

2. The method according to claim 1, characterized in that, The step of determining a terrain influence factor for each grid cell includes: For each grid cell, based on the surface topography data, a fluid accumulation amount and a surface slope value are obtained; Based on the fluid accumulation and surface slope value of the grid cell, the terrain influence factor of the grid cell is obtained.

3. The method according to claim 2, characterized in that, The terrain influence factor is obtained using the following formula: ； in, Let be the terrain influence factor of the grid cell in the i-th row and j-th column. The fluid accumulation amount for that grid cell. The surface slope value of the grid cell is given by k, m, and n, which are process coefficients associated with the chemical mechanical polishing process parameters.

4. The method according to claim 1, characterized in that, The step of generating an initial propagation contamination value includes: Construct a weight kernel, the value of each element in the weight kernel being determined by the terrain influence factor of the neighboring grid cell; The initial propagation pollution value is obtained by performing a convolution operation on the surface heavy metal concentration distribution data of the grid cell and its neighboring grid cells using the weight kernel.

5. The method according to claim 1, characterized in that, The step of extracting a global contamination index characterizing the overall contamination level of the wafer from the initial contamination distribution map includes: Statistically analyze the numerical distribution of all initial propagation pollution values ​​in the initial pollution distribution map. The values ​​at preset quantiles of the numerical distribution are obtained as the global contamination index.

6. The method according to any one of claims 1-5, characterized in that, The method further includes: After generating the initial pollution distribution map, the initial pollution distribution map is corrected to obtain a corrected pollution distribution map; The step of extracting a global contamination index characterizing the overall contamination level of the wafer from the initial contamination distribution map specifically includes: The global pollution index is extracted from the modified pollution distribution map.

7. The method according to claim 6, characterized in that, The step of correcting the initial pollution distribution map includes: Obtain the structural layout information of the through-silicon vias on the wafer to be analyzed; Based on the structural layout information, a via proximity factor is determined for each grid cell; Based on the initial propagation contamination value and the via proximity factor for each grid cell, a corrected contamination value is generated, and the corrected contamination values ​​of all grid cells constitute the corrected contamination distribution map.

8. The method according to claim 7, characterized in that, The via proximity factor characterizes the distance relationship between the grid cell and the nearest via edge.

9. A heavy metal residue detection and analysis system for TSV wafers after CMP cleaning, characterized in that, The system includes: The data acquisition module is used to acquire surface morphology data and surface heavy metal concentration distribution data of the wafer to be analyzed; The influence factor determination module is used to divide the surface of the wafer to be analyzed into multiple grid cells of a preset size based on the surface topography data, and to determine a topography influence factor for each grid cell. An initial value generation module is used to generate an initial propagation pollution value for each grid cell based on the surface heavy metal concentration distribution data of the grid cell and its neighboring grid cells and the terrain influence factor of the neighboring grid cells. The distribution map generation module is used to generate an initial contamination distribution map of the wafer to be analyzed based on the initial propagation contamination values ​​of all grid cells. The indicator extraction module is used to extract a global contamination indicator that characterizes the overall contamination level of the wafer from the initial contamination distribution map.

10. The system according to claim 9, characterized in that, Also includes: The correction module is used to correct the initial pollution distribution map after the distribution map generation module generates the initial pollution distribution map, so as to obtain a corrected pollution distribution map. Specifically, the indicator extraction module is used for: The global pollution index is extracted from the modified pollution distribution map.

Images