Correlation analysis method and device based on parameter-level defect data

Abstract

The invention provides a correlation analysis method and device based on parameter-level defect data. The correlation analysis method comprises the steps of obtaining electrical parameter-level data and a first data table when a wafer is tested; obtaining container-level data and a second data table of wafer testing; combining the first data table and the second data table to generate a third data table; constructing at least one fourth data table, acquiring field data of any to-be-associated container-level parameter field from the third data table, and adding the field data to any fourth data table; obtaining at least one variance goodness of fit based on the to-be-associated electrical parameter field and the to-be-associated container-level parameter field in each fourth data table; and determining a sorting result of the at least one variance goodness of fit according to a preset value. In this way, the result accuracy of correlation analysis between the electrical parameter level data and the container level parameter data can be improved through the determined sorting result of the at least one variance goodness of fit.

Description

Technical Field

[0001] This application relates to the field of semiconductor technology, and in particular to a correlation analysis method and apparatus based on parameter-level defect data. Background Technology

[0002] Typically, in chip manufacturing, wafers are first processed, then subjected to Wafer Acceptance Test (WAT) and chip probing (CP). Based on the WAT and CP test results, the wafers are then processed again. This alternating processing and testing process is used to fabricate chips from wafers. During wafer testing, WAT and CP can be performed sequentially. Correlation analysis can then be performed on the WAT and CP test results. Based on the WAT test results and the correlation analysis results, the wafers can be processed again to improve wafer yield.

[0003] Generally, trend charts, scatter plots, or box plots can be used to visually compare WAT test results (electrical parameter level data) and CP test results (capacitor (Bin) level parameter data) to identify potential correlations between the two. For example, analyzing the transistor saturation voltage (NMOSVtsat) in the electrical parameter level data and the threshold voltage distribution region (Vccmin Limited Yield) in the Bin level data can be performed. Figure 1 As shown, by observing the histogram of NMOS Vtsat and the data distribution trend of Vccmin Limited Yield, it can be determined that NMOS Vtsat and Vccmin Limited Yield are correlated, i.e., as NMOS Vtsat increases, Vccmin Limited Yield shows a significant decreasing trend.

[0004] On the one hand, chart analysis is used to make a direct comparison between the two sets of sample data (electrical parameter level data and Bin level parameter data). However, because there may be abnormal noise in the sample data caused by process abnormalities, it may be difficult for operators to observe whether the two sets of sample data are correlated from the chart.

[0005] On the other hand, process anomalies are low-probability events. When the sample size of process anomalies is small, there is insufficient effective information and it is difficult to trace the source. When the sample size of process anomalies is large enough, various process anomalies with different mechanisms are mixed together. In the process of identification, it is necessary to find typical process anomaly groups and then analyze each process anomaly group one by one, which may consume a lot of time, manpower and resources.

[0006] In related technologies, images can be used to display potential correlations between response variables and factors. Firstly, by observing whether changes in electrical parameter-level data exhibit continuous and gradual image characteristics in their spatial distribution, a preliminary judgment can be made as to whether changes in electrical parameter-level data can cause the failure of Bin-level parameter data. Figure 2 As shown, observe from right to left. Figure 2 (d) Figure 2 (c) Figure 2 (b) and Figure 2 In (a), as NMOSVtsat increases, the number of failure data points in the electrical parameter level data continuously increases, a gradual process that can be observed in the Bin-level parameter data graph. Then, correlation analysis is performed between the corresponding Bin-level parameter data (whose distribution area can be represented by BinYield) and device parameters (e.g., transistor saturation drain current, threshold voltage, off-state current, metal wire resistance, and chip resistance), yielding results such as... Figure 3 The correlations shown.

[0007] In related technologies, wafers can be reasonably classified using cluster analysis based on the spatial similarity of Bin-level parameter data; then, further correlation analysis can be performed on the wafers classified into each category using graphical methods. However, the above correlation analysis methods rely on past human experience for analysis and judgment, resulting in high learning costs, low efficiency, and significant interference, making it difficult to guarantee the accuracy of the correlation analysis results. Summary of the Invention

[0008] This application provides a correlation analysis method and apparatus based on parameter-level defect data, which can solve the problem of low accuracy of correlation analysis results in the prior art.

[0009] In a first aspect, a correlation analysis method based on parameter-level defect data is provided, comprising: acquiring electrical parameter-level data of a wafer undergoing testing, and acquiring a first data table according to a first preset rule and the electrical parameter-level data, wherein the first data table includes at least one first identifier field and at least one electrical parameter field to be associated; acquiring container-level parameter data of the wafer undergoing testing, and acquiring a second data table according to a second preset rule and the container-level parameter data, wherein the second data table includes at least one second identifier field and at least one container-level parameter field to be associated; merging the first data table and the second data table according to the same identifier field among the at least one first identifier field and the at least one second identifier field. A third data table is generated; using the field data corresponding to any of the same identifier fields as row identifiers and the field data corresponding to any at least one electrical parameter field to be associated as column identifiers, at least one fourth data table is constructed, and according to the column identifiers and the row identifiers, the field data of any container-level parameter field to be associated is obtained from the third data table and added to the fourth data table; based on the electrical parameter field to be associated and the container-level parameter field to be associated in each fourth data table, at least one variance goodness of fit is obtained; the ranking result of the at least one variance goodness of fit is determined according to a preset value, and the ranking result is used to represent the correlation between the electrical parameter field to be associated and the container-level parameter field to be associated.

[0010] In some embodiments of this application, obtaining at least one variance goodness of fit based on the electrical parameter field to be associated and the container-level parameter field to be associated in each of the fourth data tables includes: determining each column of data in each of the fourth data tables as a group of container-level parameter field data to be associated according to the column identifier; obtaining the within-group sum of squares, component sum of squares, and total sum of squares of the container-level parameter field data to be associated; and obtaining the variance goodness of fit for each of the fourth data tables based on the within-group sum of squares, the component sum of squares, and the total sum of squares.

[0011] In some embodiments of this application, the step of determining each column of data in the fourth data table as a group of container-level parameter field data to be associated based on the column identifier, and obtaining the within-group sum of squares, between-group sum of squares, and total sum of squares of the container-level parameter field data to be associated, includes: determining each column of data in the fourth data table as a group of container-level parameter field data to be associated based on the column identifier, obtaining the group data mean of each group of container-level parameter field data to be associated; obtaining the total data mean of the fourth data table based on the group data mean and the number of field data of the electrical parameter field to be associated in the fourth data table; obtaining the within-group sum of squares based on the group data mean and the total data mean; obtaining the between-group sum of squares based on the field data of the electrical parameter field to be associated in the fourth data table and the group data mean; and obtaining the total sum of squares based on the within-group sum of squares and the between-group sum of squares.

[0012] In some embodiments of this application, after obtaining at least one variance goodness of fit based on the electrical parameter field to be associated and the container-level parameter field to be associated in each of the fourth data tables, the method further includes: obtaining an optimized variance goodness of fit based on the number of fields of the electrical parameter field to be associated in the fourth data table, the number of data items of the field data of the electrical parameter field to be associated in the fourth data table, and the variance goodness of fit; after determining the sorting result of the at least one variance goodness of fit based on a preset value, the method further includes: sorting the variance goodness of fit with the same value in the sorting result again based on the optimized variance goodness of fit.

[0013] In some embodiments of this application, the step of re-sorting based on the optimized variance fit goodness includes: determining an optimized sorting result for the optimized variance fit goodness, wherein the optimized sorting result is obtained by arranging the absolute values ​​of the differences between the optimized variance fit goodness and the preset value in ascending order; and updating the sorting result based on the optimized sorting result.

[0014] In some embodiments of this application, the first preset rule includes a first data preprocessing rule and a wafer acceptance test field filtering rule. The first data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the electrical parameter level data. The wafer acceptance test field filtering rule is used to filter out the at least one electrical parameter field to be associated from the electrical parameter level data.

[0015] In some embodiments of this application, the second preset rule includes a second data preprocessing rule and a wafer test field filtering rule. The second data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the container-level parameter data. The wafer test field filtering rule is used to filter out at least one container-level parameter field to be associated from the container-level parameter data.

[0016] In some embodiments of this application, the electrical parameter field to be associated includes at least one of the following: saturation current, off-state current, and chip resistance; the capacitor-level parameter field to be associated includes at least one of the following: threshold voltage, on-resistance, and source-drain breakdown current.

[0017] Secondly, a correlation analysis method and apparatus based on parameter-level defect data is provided, comprising: a first acquisition module, a second acquisition module, a generation module, a processing module, a third acquisition module, and a determination module, wherein...

[0018] The first acquisition module is used to acquire electrical parameter-level data of the wafer undergoing testing, and acquire a first data table according to a first preset rule and the electrical parameter-level data. The first data table includes at least one first identifier field and at least one electrical parameter field to be associated. The second acquisition module is used to acquire container-level parameter data of the wafer undergoing testing, and acquire a second data table according to a second preset rule and the container-level parameter data. The second data table includes at least one second identifier field and at least one container-level parameter field to be associated. The generation module is used to merge the first data table and the second data table according to the same identifier field in the at least one first identifier field and the at least one second identifier field to generate a third data table. The processing module... The third acquisition module is used to construct at least one fourth data table, using the field data corresponding to any of the same identifier fields as row identifiers and the field data corresponding to the at least one electrical parameter field to be associated as column identifiers. Based on the column identifiers and row identifiers, the module retrieves the field data of any container-level parameter field to be associated from the third data table and adds it to any of the fourth data tables. The third acquisition module is used to obtain at least one variance goodness of fit based on the electrical parameter field to be associated and the container-level parameter field to be associated in each of the fourth data tables. The determining module is used to determine the ranking result of the at least one variance goodness of fit based on a preset value. The ranking result represents the degree of correlation between the electrical parameter field to be associated and the container-level parameter field to be associated.

[0019] In some embodiments of this application, the third acquisition module is used to: determine each column of data in each of the fourth data tables as a group of container-level parameter field data to be associated based on the column identifier; acquire the within-group sum of squares, component sum of squares, and total sum of squares of the container-level parameter field data to be associated; and acquire the variance goodness of fit of each of the fourth data tables based on the within-group sum of squares, component sum of squares, and total sum of squares.

[0020] In some embodiments of this application, the third acquisition module is configured to: determine each column of data in the fourth data table as a group of container-level parameter field data to be associated based on the column identifier; acquire the group data mean of each group of container-level parameter field data to be associated; acquire the total data mean of the fourth data table based on the group data mean and the number of field data of the electrical parameter field to be associated in the fourth data table; acquire the sum of squared deviations within the group based on the group data mean and the total data mean; acquire the sum of squared deviations between groups based on the field data of the electrical parameter field to be associated in the fourth data table and the group data mean; and obtain the total sum of squared deviations based on the sum of squared deviations within the group and the sum of squared deviations between the group.

[0021] In some embodiments of this application, the apparatus further includes: a sorting module;

[0022] The third acquisition module is further configured to: after acquiring at least one variance fit goodness based on the electrical parameter field to be associated and the container-level parameter field to be associated in each of the fourth data tables, obtain an optimized variance fit goodness based on the number of fields of the electrical parameter field to be associated in the fourth data table, the number of data items of the field data of the electrical parameter field to be associated in the fourth data table, and the variance fit goodness.

[0023] The sorting module is used to, after determining the sorting result of at least one variance fit goodness of fit based on a preset value, re-sort the variance fit goodness of fit results with the same value based on the optimized variance fit goodness of fit.

[0024] In some embodiments of this application, the sorting module is used to: determine the optimized sorting result of the optimized variance fit, wherein the optimized sorting result is obtained by arranging the absolute values ​​of the difference between the optimized variance fit and the preset value in ascending order; and update the sorting result according to the optimized sorting result.

[0025] In some embodiments of this application, the first preset rule includes a first data preprocessing rule and a wafer acceptance test field filtering rule. The first data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the electrical parameter level data. The wafer acceptance test field filtering rule is used to filter out the at least one electrical parameter field to be associated from the electrical parameter level data.

[0026] In some embodiments of this application, the second preset rule includes a second data preprocessing rule and a wafer test field filtering rule. The second data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the container-level parameter data. The wafer test field filtering rule is used to filter out at least one container-level parameter field to be associated from the container-level parameter data.

[0027] In some embodiments of this application, the electrical parameter field to be associated includes at least one of the following: saturation current, off-state current, and chip resistance; the capacitor-level parameter field to be associated includes at least one of the following: threshold voltage, on-resistance, and source-drain breakdown current.

[0028] The technical solution provided in this application first obtains electrical parameter-level data of the wafer undergoing testing, and then obtains a first data table based on a first preset rule and the electrical parameter-level data. Next, it obtains container-level parameter data of the wafer undergoing testing, and then obtains a second data table based on a second preset rule and the container-level parameter data. Next, it merges the first and second data tables based on the same identifier field in at least one first identifier field and at least one second identifier field to generate a third data table. Next, it converts the third data table into at least one fourth data table, and then obtains at least one variance goodness-of-fit based on the electrical parameter field and container-level parameter field to be associated in each fourth data table. Finally, it determines the ranking result of at least one variance goodness-of-fit based on preset values, wherein the ranking result is used to represent the degree of correlation between the electrical parameter field and the container-level parameter field to be associated. Thus, by determining the correlation between electrical parameter-level data and container-level parameter data based on the variance goodness-of-fit of parameter-level defect data, engineers can avoid the need to organize complex data and perform correlation analysis using comparative charts. Ranking the results based on at least one variance goodness of fit to determine the correlation between electrical parameter-level data and container-level parameter data in the semiconductor manufacturing stage can improve the accuracy and efficiency of correlation analysis results acquisition. Furthermore, it allows for automatic data tracing for calculations, significantly reducing anomaly analysis time and minimizing losses due to defects. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a schematic diagram illustrating the trend between the histogram of NMOS Vtsat and Vccmin Limited Yield according to some embodiments of this application;

[0031] Figure 2 This is a schematic diagram illustrating the variations of the CP Binmap according to some embodiments of this application;

[0032] Figure 3 This is a schematic diagram illustrating the relationship between WAT Parameter and Bin Yield according to some embodiments of this application;

[0033] Figure 4 This is a flowchart illustrating a correlation analysis method based on parameter-level defect data according to some embodiments of this application;

[0034] Figure 5 This is a flowchart illustrating a method for obtaining variance goodness of fit according to some embodiments of this application;

[0035] Figure 6 This is a block diagram of a correlation analysis apparatus based on parameter-level defect data, according to some embodiments of this application;

[0036] Figure 7 These are schematic block diagrams of an electronic device according to some embodiments of this application. Detailed Implementation

[0037] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0038] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in sequences other than those illustrated or described herein.

[0039] Furthermore, the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion, such that a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0040] As mentioned above, in related technologies, wafers can be reasonably classified using cluster analysis based on the spatial similarity of Bin-level parameter data; then, further correlation analysis can be performed on the wafers classified into each category using graphical methods. However, the above correlation analysis methods require analysis and judgment based on past human experience, resulting in high learning analysis costs, low efficiency, and significant human interference, making it difficult to guarantee the accuracy of the correlation analysis results.

[0041] To address the aforementioned technical problems, this application provides a correlation analysis method based on parameter-level defect data. The following description, in conjunction with the accompanying drawings, uses a correlation analysis device based on parameter-level defect data as the executing entity to illustrate the correlation analysis method based on parameter-level defect data provided in this application.

[0042] like Figure 4 As shown, the correlation analysis method based on parameter-level defect data provided in this application embodiment may include steps S401 to S406.

[0043] Step S401: Obtain electrical parameter level data of the wafer to be tested, and obtain a first data table according to the first preset rule and the electrical parameter level data.

[0044] The first data table includes at least one first identifier field and at least one electrical parameter field to be associated. The electrical parameter field to be associated includes at least one of the following: saturation current, off-state current, and sheet resistance.

[0045] In this embodiment, electrical parameter-level data refers to data obtained through wafer testing, used to monitor the normality and stability of each process step via electrical parameters. The associated electrical parameter fields correspond to the electrical parameter-level data. The electrical parameter-level data can be historical data from the operation of the semiconductor equipment.

[0046] In this embodiment of the application, the first preset rule includes a first data preprocessing rule and a wafer acceptance test field filtering rule. The first data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the electrical parameter level data. The wafer acceptance test field filtering rule is used to filter out at least one electrical parameter field to be associated from the electrical parameter level data.

[0047] For example, as shown in Table 1, the fields in the first data table include: ProductFamily (product type), TestType (test model), LotID (lot number), WaferID (wafer number), ParameterName (parameter name), and ParamValue (parameter value). The first identifier field can be: ProductFamily, TestType, LotID, and WaferID. The field name of the electrical parameter field to be associated is Parameter, and the field value of the electrical parameter field to be associated is ParamValue.

[0048] Table 1 First Data Table

[0049] ProductFamily TestType LotID WaferID Parameter ParamValue A B C0001.01 C0002.01 V 11.4 A B C0001.01 C0002.01 V 11.4 A B C0001.02 C0002.02 V 11.8 A B C0001.02 C0002.02 V 12.5 A B C0001.02 C0002.02 V 12.1 A B C0001.03 C0002.03 V 12.1 A B C0001.04 C0002.04 V 12.5 A B C0001.04 C0002.04 V 12.5 A B C0001.05 C0002.05 V 11.8 A B C0001.05 C0002.05 V 13.5

[0050] Step S402: Obtain wafer test container-level parameter data, and obtain a second data table based on the second preset rule and the container-level parameter data.

[0051] The second data table includes at least one second identifier field and at least one container-level parameter field to be associated. The container-level parameter field to be associated includes at least one of the following: threshold voltage, on-resistance, and source-drain breakdown current.

[0052] In this embodiment, container-level parameter data refers to the yield of the test data corresponding to the container-level parameter field to be associated, obtained through CP testing, i.e., defect data. The purpose of statistically analyzing container-level parameter data is to identify defective chip dies, reduce packaging and testing costs, and thus determine the wafer yield. The test data corresponding to the container-level parameter field to be associated can be historical data from the operation of semiconductor equipment.

[0053] In this embodiment of the application, the second preset rule includes a second data preprocessing rule and a wafer test field filtering rule. The second data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the container-level parameter data. The wafer test field filtering rule is used to filter out at least one container-level parameter field to be associated from the container-level parameter data.

[0054] For example, as shown in Table 2, the fields in the second data table include: ProductFamily, TestType, LotSrc, LotID, ProductID, WaferID, WaferNo, Bin, and BinYield. The second identifier field includes ProductFamily, TestType, LotID, and WaferID. At least one container-level parameter field to be associated has the field name "Container Bin," and at least one container-level parameter field to be associated has the field value "Container Yield BinYield."

[0055] Table 2 Second Data Table

[0056] ProductFamily TestType LotID ProductID WaferID Bin BinYield A B C0001.01 A C0002.01 EF 0.001 A B C0001.01 A C0002.01 EF 0.002 A B C0001.02 A C0002.02 EF 0.001 A B C0001.02 A C0002.02 EF 0.004 A B C0001.02 A C0002.02 EF 0.002 A B C0001.03 A C0002.03 EF 0.001 A B C0001.04 A C0002.04 EF 0.003 A B C0001.04 A C0002.04 EF 0.002 A B C0001.05 A C0002.05 EF 0.001 A B C0001.05 A C0002.05 EF 0.004

[0057] Step S403: Based on the same identifier field in at least one first identifier field and at least one second identifier field, merge the first data table and the second data table to generate a third data table.

[0058] In this embodiment, the first data table and the second data table are associated using the same identifier field. First, the table fields in the first data table and the second data table are obtained. Then, at least one first identifier field is extracted from the first data table, and at least one second identifier field is extracted from the second data table. Next, at least one first identifier field and at least one second identifier field are compared, and the same identifier field is determined. Finally, based on the same identifier field, the first data table and the second data table are merged to generate a third data table.

[0059] For example, as shown in Table 3, the common identifier fields in the third data table include: ProductFamily, TestType, LotID, and WaferID. The field name for the electrical parameter field to be associated in the third data table is Parameter, and the field value is ParamValue. The field name for the container-level parameter field to be associated in the third data table is ContainerBin, and at least one of the container-level parameter fields to be associated has a field value of ContainerYield.

[0060] Table 3 Third Data Table

[0061]

[0062] Step S404: Using the field data corresponding to any identical identifier field as the row identifier and the field data corresponding to any at least one electrical parameter field to be associated as the column identifier, construct at least one fourth data table, and according to the column identifier and row identifier, obtain the field data of any container-level parameter field to be associated from the third data table and add it to the fourth data table.

[0063] In this embodiment, the third data table is logically transformed to generate a fourth data table. The number of fields in the fourth data table is the product of the number of fields in at least one electrical parameter field to be associated and the number of fields in at least one container-level parameter field to be associated.

[0064] For example, as shown in Table 4, the field data corresponding to any one or more electrical parameter fields to be associated in the third data table includes 11.4, 11.8, 12.1, 12.5, and 13.5. The column identifiers of the fourth data table are determined to be 11.4, 11.8, 12.1, 12.5, and 13.5. The field data corresponding to LotID from the fields with the same identifier, including ProductFamily, TestType, LotID, and WaferID, is selected as the row identifier of the fourth data table. Based on the field data corresponding to the LotID field and the field data corresponding to the electrical parameter fields to be associated, the corresponding BinYield data is found in the third data table and added to the fourth data table. It should be noted that empty data in the fourth data table is identified using a preset character, which can be NULL.

[0065] Table 4, Fourth Data Table

[0066]

[0067] Step S405: Based on the electrical parameter field to be associated and the container-level parameter field to be associated in each fourth data table, obtain at least one variance goodness of fit.

[0068] In the embodiments of this application, the variance analysis method can be used to calculate the difference of the data in the fourth data table in order to obtain the variance goodness of fit of each fourth data table.

[0069] In some possible implementations, such as Figure 5 As shown in the embodiment of this application, step S405 includes steps S4051 and S4052.

[0070] Step S4051: Based on the column identifier, determine each column of data in each fourth data table as a group of container-level parameter field data to be associated, and obtain the within-group sum of squares, component sum of squares, and total sum of squares of the container-level parameter field data to be associated.

[0071] Specifically, step S4051 includes: determining each column of data in the fourth data table as a group of container-level parameter field data to be associated based on the column identifier, and obtaining the group data mean of each group of container-level parameter field data to be associated; obtaining the total data mean of the fourth data table based on the group data mean and the number of data points of the field data of the electrical parameter field to be associated in the fourth data table; obtaining the sum of squared deviations within the group based on the group data mean and the total data mean; obtaining the sum of squared deviations between groups based on the field data of the electrical parameter field to be associated in the fourth data table and the group data mean; and obtaining the total sum of squared deviations based on the sum of squared deviations within the group and the sum of squared deviations between the groups.

[0072] In this embodiment of the application, the formula for the mean of group data is... Where i represents the column containing the field data of the container-level parameter field to be associated, K represents the number of valid LotIDs (LotIDs with BinYield data), and y is the data value of BinYield.

[0073] For example, the mean of the group data is obtained according to Table 4, specifically including: the BinYield corresponding to the first group of container-level parameter field data (11.4) includes 0.01 and 0.01, and the mean of the first group of data is... The BinYield values ​​for the second set of container-level parameter field data (11.8) are 0.01 and 0.04, respectively. The mean of the second set of data is... The BinYield values ​​for the third group of container-level parameter field data (12.1) include 0.02 and 0.02, and the mean of the third group of data is... The BinYield values ​​for the fourth group of container-level parameter field data (12.5) are 0.04, 0.03, and 0.01, respectively. The mean of the fourth group of data is... The BinYield corresponding to the fifth group of container-level parameter field data (13.5) is 0.02, and the mean of the fifth group of data is...

[0074] In this application embodiment, the formula for the average of total data is... Where n represents the number of data items in the BinYield field, and y represents the data value of BinYield.

[0075] For example, based on the example above, the total data mean,

[0076]

[0077] In the embodiments of this application, the formula for the sum of squared deviations between groups is... Where m is the number of field data in the container-level parameter field to be associated, that is, the number of group data averages.

[0078] For example, the sum of squared deviations between groups

[0079] In the embodiments of this application, the formula for the sum of squared deviations within a group is... For example, within-group sum of squares

[0080] In this embodiment, the total sum of squared deviations (SST) is calculated as SSR + SSE. For example, the total sum of squared deviations (SST) is 3.52 * 10^- ... -6 +9.18*10-6 =1.27*10 -5 .

[0081] It should be noted that the between-group sum of squares is used to represent the level variance, which is a measure of the difference between samples from different populations, that is, it represents the error in the detection parameter values ​​at different WAT stages under different LOTs. The within-group sum of squares, on the other hand, measures the variance of sample data within the same population, that is, it represents the difference in the detection parameter values ​​at the same WAT stage under different LOTs.

[0082] Step S4052: Based on the sum of squared deviations within the group, the sum of squared deviations of the components, and the total sum of squared deviations, obtain the variance goodness of fit of the electrical parameter field to be associated and the container-level parameter field to be associated in each fourth data table.

[0083] In the embodiments of this application, the variance fit goodness formula For example, based on the above example, variance goodness of fit

[0084] Step S406: Determine at least one ranking result of variance fit based on preset values.

[0085] The sorting result is used to indicate the degree of correlation between the electrical parameter field to be associated and the container-level parameter field to be associated.

[0086] In this embodiment of the application, after obtaining the variance goodness of fit for each fourth data table, the absolute values ​​of the differences between the variance goodness of fit and preset values ​​are sorted from smallest to largest, and at least one goodness of fit ranking result is determined based on this. For example, the preset values ​​can be 1, 2, 10, and 100, etc.

[0087] It is understandable that if the variance goodness of fit is closer to 1, it indicates that the correlation between the electrical parameter field to be correlated and the container-level parameter field to be correlated is higher.

[0088] In the sorting results, the container-level parameter field to be associated can also include the test type TestType and the container name Bin name, and the electrical parameter field to be associated can also include the parameter name Param and the parameter number ParamNo.

[0089] In some embodiments, after step S405, the following can be performed: obtain the optimized variance fit based on the number of fields of the electrical parameter field to be associated in the fourth data table, the number of data in the field data of the electrical parameter field to be associated in the fourth data table, and the variance fit; after step S406, the following can be performed: sort the variance fit results with the same value again based on the optimized variance fit.

[0090] The process of re-sorting based on the optimized variance fit includes: determining the optimized variance fit ranking result by taking the absolute values ​​of the differences between the optimized variance fit and the preset values ​​from smallest to largest; and updating the ranking result based on the optimized ranking result.

[0091] As an example, the relevance ranking results are shown in Table 5. Wherein, R... 2 For variance, goodness of fit To optimize the variance fit.

[0092] Table 5. Correlation Ranking Results

[0093]

[0094]

[0095] It is understandable that the goodness of fit of variance represents the degree to which the regression line fits the observed values ​​(the Bin Yield of each Lot under a certain measurement parameter). If more independent variables are introduced, the goodness of fit of variance will become larger and closer to 1. This is because the adjusted goodness of fit of variance takes into account both the sample size n and the number of independent variables m, so it is more meaningful when there are many independent variables.

[0096] The correlation analysis method based on parameter-level defect data provided in this application first acquires electrical parameter-level data of the wafer undergoing testing, and obtains a first data table according to a first preset rule and the electrical parameter-level data. Next, it acquires container-level parameter data of the wafer undergoing testing, and obtains a second data table according to a second preset rule and the container-level parameter data. Next, it merges the first and second data tables according to the same identifier field in at least one first identifier field and at least one second identifier field to generate a third data table. Next, it converts the third data table into at least one fourth data table, and then obtains at least one variance goodness-of-fit based on the electrical parameter field and container-level parameter field to be associated in each fourth data table. Finally, it determines the ranking result of at least one variance goodness-of-fit based on preset values, wherein the ranking result is used to represent the degree of correlation between the electrical parameter field and the container-level parameter field to be associated. Thus, by determining the correlation between electrical parameter-level data and container-level parameter data based on the variance goodness-of-fit of parameter-level defect data, engineers can avoid the need for tedious data processing and comparative analysis charts for correlation analysis. Ranking the results based on at least one variance goodness of fit to determine the correlation between electrical parameter-level data and container-level parameter data in the semiconductor manufacturing stage can improve the accuracy and efficiency of correlation analysis results acquisition. Furthermore, it allows for automatic data tracing for calculations, significantly reducing anomaly analysis time and minimizing losses due to defects.

[0097] like Figure 6 As shown in the figure, this application embodiment also provides a correlation analysis device based on parameter-level defect data. The correlation analysis device based on parameter-level defect data may include a first acquisition module 61, a second acquisition module 62, a generation module 63, a processing module 64, a third acquisition module 65, and a determination module 66.

[0098] The first acquisition module 61 is used to acquire electrical parameter level data of the wafer undergoing testing, and to acquire a first data table according to a first preset rule and the electrical parameter level data. The first data table includes at least one first identifier field and at least one electrical parameter field to be associated; for example, such as Figure 4 As shown, the first acquisition module 61 can be used to execute step S401.

[0099] The second acquisition module 62 is used to acquire container-level parameter data for wafer testing, and according to a second preset rule and the container-level parameter data, acquire a second data table. The second data table includes at least one second identifier field and at least one container-level parameter field to be associated; for example, such as... Figure 4 As shown, the second acquisition module 62 can be used to execute step S402.

[0100] The generation module 63 is used to merge the first data table and the second data table to generate a third data table based on the common identifier field in the at least one first identifier field and the at least one second identifier field; for example, such as Figure 4 As shown, the generation module 63 can be used to perform step S403.

[0101] The processing module 64 is used to construct at least one fourth data table, using the field data corresponding to any of the same identifier fields as row identifiers and the field data corresponding to any of the at least one electrical parameter field to be associated as column identifiers. Based on the column identifiers and row identifiers, it retrieves the field data of any container-level parameter field to be associated from the third data table and adds it to the fourth data table; for example, such as... Figure 4 As shown, the processing module 64 can be used to execute step S404.

[0102] The third acquisition module 65 is used to acquire at least one variance goodness of fit based on the electrical parameter field to be associated and the container-level parameter field to be associated in each of the fourth data tables; for example, Figure 4 As shown, the third acquisition module 65 can be used to execute step S405.

[0103] The determining module 66 is used to determine the ranking result of the at least one variance goodness of fit based on a preset value. The ranking result is used to represent the correlation between the electrical parameter field to be correlated and the container-level parameter field to be correlated. For example, Figure 4 As shown, the determination module 66 can be used to perform step S406.

[0104] In some embodiments of this application, such as Figure 6 As shown, the third acquisition module 65 is used to: determine each column of data in each of the fourth data tables as a group of container-level parameter field data to be associated according to the column identifier; acquire the within-group sum of squares, component sum of squares, and total sum of squares of the container-level parameter field data to be associated; and acquire the variance goodness of fit of each of the fourth data tables according to the within-group sum of squares, component sum of squares, and total sum of squares.

[0105] In some embodiments of this application, such as Figure 6 As shown, the third acquisition module 65 is used to: determine each column of data in the fourth data table as a group of container-level parameter field data to be associated according to the column identifier; obtain the group data mean of each group of container-level parameter field data to be associated; obtain the total data mean of the fourth data table according to the group data mean and the number of field data of the electrical parameter field to be associated in the fourth data table; obtain the sum of squares of deviations within the group according to the group data mean and the total data mean; obtain the sum of squares of deviations between groups according to the field data of the electrical parameter field to be associated in the fourth data table and the group data mean; and obtain the total sum of squares of deviations according to the sum of squares of deviations within the group and the sum of squares of deviations between groups.

[0106] In some embodiments of this application, such as Figure 6 As shown, the device further includes: a sorting module 67; the third acquisition module 65 is further configured to: after acquiring at least one variance fit value based on the electrical parameter field to be associated and the container-level parameter field to be associated in each of the fourth data tables, acquire an optimized variance fit value based on the number of fields of the electrical parameter field to be associated in the fourth data table, the number of data items of the field data of the electrical parameter field to be associated in the fourth data table, and the variance fit value;

[0107] The sorting module 67 is used by the determining module 66 to obtain at least one variance fit value based on the electrical parameter field to be associated and the container-level parameter field to be associated in each of the fourth data tables, and then to sort the variance fit values ​​with the same value in the sorting results again according to the optimized variance fit value.

[0108] In some embodiments of this application, such as Figure 6 As shown, the sorting module 67 is used to: determine the optimized sorting result of the optimized variance fit, wherein the optimized sorting result is obtained by arranging the absolute values ​​of the difference between the optimized variance fit and the preset value in ascending order; and update the sorting result according to the optimized sorting result.

[0109] In some embodiments of this application, the first preset rule includes a first data preprocessing rule and a wafer acceptance test field filtering rule. The first data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the electrical parameter level data. The wafer acceptance test field filtering rule is used to filter out the at least one electrical parameter field to be associated from the electrical parameter level data.

[0110] In some embodiments of this application, the second preset rule includes a second data preprocessing rule and a wafer test field filtering rule. The second data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the container-level parameter data. The wafer test field filtering rule is used to filter out at least one container-level parameter field to be associated from the container-level parameter data.

[0111] In some embodiments of this application, the electrical parameter field to be associated includes at least one of the following: saturation current, off-state current, and chip resistance; the capacitor-level parameter field to be associated includes at least one of the following: threshold voltage, on-resistance, and source-drain breakdown current.

[0112] Regarding the apparatus in the above embodiments, the specific manner in which each unit performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0113] It is understood that, in conjunction with the various examples described in the embodiments of this application, those skilled in the art should readily recognize that this application can be implemented in the form of hardware, computer software, or a combination of hardware and software (hardware and computer software). To achieve the above functions, the correlation analysis apparatus based on parameter-level defect data includes at least one of the hardware structure and software module corresponding to each function. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0114] This application embodiment can divide the correlation analysis device based on parameter-level defect data into functional units according to the above method example. For example, each function can be divided into a separate functional unit, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or as a software functional unit. It should be noted that the unit division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0115] Figure 7 This is a schematic diagram of the structure of an electronic device provided in this application. For example... Figure 7 As shown, the electronic device may include a processor 71 and a memory 72 for storing executable instructions of the processor 71; wherein the processor 71 is configured to execute the instructions to implement the correlation analysis method based on parameter-level defect data in the above embodiment.

[0116] In addition, the electronic device may also include a communication bus 73 and at least one communication interface 74.

[0117] The processor 71 may be a central processing unit (CPU), a microprocessor unit, an application-specific integrated circuit (ASIC), or one or more integrated circuits used to control the execution of the program of the present application.

[0118] The communication bus 73 is a signal path used to transmit information between the aforementioned components.

[0119] Communication interface 74 uses any transceiver-like device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0120] The memory 72 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital versatile optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 72 may exist independently and be connected to the processor 71 via a communication bus 73. The memory 72 may also be integrated with the processor 71.

[0121] The memory 72 stores instructions for executing the scheme of this application, and the processor 71 controls the execution. The processor 71 executes the program or instructions stored in the memory 72 to realize the functions of the method of this application.

[0122] As an example, combined Figure 6 The functions implemented by the first acquisition module 61, the second acquisition module 62, the generation module 63, the processing module 64, the third acquisition module 65, and the determination module 66 in the correlation analysis device based on parameter-level defect data are the same as those of the first acquisition module 61, the second acquisition module 62, the generation module 63, the processing module 64, the third acquisition module 65, and the determination module 66. Figure 7 The processor 71 in it has the same function.

[0123] In a specific implementation, as one embodiment, the electronic device may further include an output device 75 and an input device 76. The output device 75 communicates with the processor 71 and can display information in various ways. For example, the output device 75 may be a liquid crystal display (LCD), a light-emitting diode (LED) display device, a cathode ray tube (CRT) display device, or a projector, etc. The input device 76 communicates with the processor 71 and can accept user input in various ways. For example, the input device 76 may be a mouse, keyboard, touchscreen device, or sensing device, etc.

[0124] Those skilled in the art will understand that Figure 7The structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or use different component arrangements. Figure 7 The electronic devices in the system can be servers, clients, or other devices.

[0125] In addition, this application also provides a computer-readable storage medium storing a program or instructions. When the instructions in the computer-readable storage medium are executed by a processor, the electronic device is able to perform the correlation analysis method based on parameter-level defect data provided in the above embodiments. Optionally, the readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, or optical data storage device, etc.

[0126] In addition, this application also provides a computer program product, including a computer program / instructions, which is stored in a non-volatile readable storage medium. When the computer program product is executed by at least one processor, it causes an electronic device to perform the correlation analysis method based on parameter-level defect data as provided in the above embodiments.

[0127] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the claims.

[0128] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A correlation analysis method based on parameter-level defect data, characterized in that, include: Obtain electrical parameter level data of the wafer to be tested, and obtain a first data table according to a first preset rule and the electrical parameter level data. The first data table includes at least one first identifier field and at least one electrical parameter field to be associated. Obtain container-level parameter data for wafer testing, and obtain a second data table based on a second preset rule and the container-level parameter data. The second data table includes at least one second identifier field and at least one container-level parameter field to be associated. Based on the common identifier field in at least one first identifier field and at least one second identifier field, the first data table and the second data table are merged to generate a third data table; Using the field data corresponding to any of the same identifier fields as row identifiers and the field data corresponding to any of the at least one electrical parameter fields to be associated as column identifiers, at least one fourth data table is constructed. Based on the column identifiers and the row identifiers, the field data of any container-level parameter field to be associated is obtained from the third data table and added to the fourth data table. Based on the electrical parameter field to be associated and the container-level parameter field to be associated in each of the fourth data tables, at least one variance goodness of fit is obtained. The ranking result of the at least one variance fit is determined based on a preset value, and the ranking result is used to represent the degree of correlation between the electrical parameter field to be associated and the container-level parameter field to be associated.

2. The correlation analysis method based on parameter-level defect data according to claim 1, characterized in that, The step of obtaining at least one variance goodness of fit based on the electrical parameter field and the container-level parameter field to be associated in each of the fourth data tables includes: Based on the column identifier, each column of data in each of the fourth data tables is determined as a group of container-level parameter field data to be associated, and the within-group sum of squares, component sum of squares, and total sum of squares of the container-level parameter field data to be associated are obtained; The variance goodness of fit of each of the fourth data tables is obtained based on the sum of squared deviations within the group, the sum of squared deviations of the components, and the total sum of squared deviations.

3. The correlation analysis method based on parameter-level defect data according to claim 2, characterized in that, The step of determining each column of data in the fourth data table as a group of container-level parameter field data to be associated based on the column identifier, and obtaining the within-group sum of squares, between-group sum of squares, and total sum of squares of the container-level parameter field data to be associated, includes: Based on the column identifier, each column of data in the fourth data table is determined as a group of container-level parameter field data to be associated, and the average value of each group of container-level parameter field data to be associated is obtained. Based on the average of the group data and the number of data points in the field data of the electrical parameter field to be associated in the fourth data table, obtain the total average of the data in the fourth data table; Based on the mean of the group data and the mean of the total data, obtain the sum of squared deviations within the group; Based on the field data of the electrical parameter field to be associated in the fourth data table and the mean of the group data, obtain the sum of squared deviations between groups; The total sum of squares is obtained based on the sum of squares within groups and the sum of squares between groups.

4. The correlation analysis method based on parameter-level defect data according to claim 1, characterized in that, After obtaining at least one variance goodness of fit based on the electrical parameter field and the container-level parameter field to be associated in each of the fourth data tables, the method further includes: Based on the number of fields in the electrical parameter field to be associated in the fourth data table, the number of data points in the field data of the electrical parameter field to be associated in the fourth data table, and the variance fit, the optimized variance fit is obtained. After determining the ranking result of the at least one variance fit goodness of fit based on the preset value, the method further includes: For the variance fit scores with the same value in the sorting results, the results are sorted again based on the optimized variance fit scores.

5. The correlation analysis method based on parameter-level defect data according to claim 4, characterized in that, The step of re-sorting based on the optimized variance goodness of fit includes: The optimization ranking result of the optimized variance fit is determined by arranging the absolute values ​​of the difference between the optimized variance fit and the preset value in ascending order. Update the sorting result based on the optimized sorting result.

6. The correlation analysis method based on parameter-level defect data according to claim 1, characterized in that, The first preset rule includes a first data preprocessing rule and a wafer acceptance test field filtering rule. The first data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the electrical parameter level data. The wafer acceptance test field filtering rules are used to filter out at least one electrical parameter field to be associated from the electrical parameter level data.

7. The correlation analysis method based on parameter-level defect data according to claim 1, characterized in that, The second preset rule includes a second data preprocessing rule and a wafer test field filtering rule. The second data preprocessing rule is used to remove nulls, duplicates, and invalid fields from the container-level parameter data. The wafer test field filtering rules are used to filter out at least one container-level parameter field to be associated from the container-level parameter data.

8. The correlation analysis method based on parameter-level defect data according to any one of claims 1-7, characterized in that, The electrical parameter field to be associated includes at least one of the following: saturation current, off-state current, and sheet resistance; The container-level parameter fields to be associated include at least one of the following: threshold voltage, on-resistance, and source-drain breakdown current.

9. A correlation analysis device based on parameter-level defect data, characterized in that, include: The system comprises a first acquisition module, a second acquisition module, a generation module, a processing module, a third acquisition module, and a determination module, wherein... The first acquisition module is used to acquire electrical parameter level data of the wafer undergoing testing, and to acquire a first data table according to a first preset rule and the electrical parameter level data. The first data table includes at least one first identifier field and at least one electrical parameter field to be associated. The second acquisition module is used to acquire container-level parameter data of wafer testing, and acquire a second data table according to the second preset rule and the container-level parameter data. The second data table includes at least one second identifier field and at least one container-level parameter field to be associated. The generation module is used to merge the first data table and the second data table to generate a third data table based on the same identifier field in the at least one first identifier field and the at least one second identifier field; The processing module is used to construct at least one fourth data table, using the field data corresponding to any of the same identifier fields as row identifiers and the field data corresponding to any of the at least one electrical parameter fields to be associated as column identifiers, and to obtain the field data of any container-level parameter field to be associated from the third data table according to the column identifiers and the row identifiers, and add it to the fourth data table. The third acquisition module is used to acquire at least one variance goodness of fit based on the electrical parameter field to be associated and the container-level parameter field to be associated in each of the fourth data tables. The determining module is used to determine the ranking result of the at least one variance fit goodness of fit based on a preset value. The ranking result is used to represent the degree of correlation between the electrical parameter field to be associated and the container-level parameter field to be associated.

10. The correlation analysis device based on parameter-level defect data according to claim 9, characterized in that, The third acquisition module is used for: Based on the column identifier, each column of data in each of the fourth data tables is determined as a group of container-level parameter field data to be associated, and the within-group sum of squares, component sum of squares, and total sum of squares of the container-level parameter field data to be associated are obtained; The variance goodness of fit of each of the fourth data tables is obtained based on the sum of squared deviations within the group, the sum of squared deviations of the components, and the total sum of squared deviations.

Images