Method and device for optimizing integrated circuit metal layer photolithography process

By training the finished product yield prediction model, identifying the combination of key process parameters and adjustment terms, optimizing the integrated circuit metal layer lithography process, solving the problems of wire breaks of metal layer and local residual aluminum dots, and improving the finished product yield of the chip metal layer.

CN118800645BActive Publication Date: 2024-11-15TIANSHUI TIANGUANG SEMICON
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411281412.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-13
Publication Date
2024-11-15
Estimated Expiration
2044-09-13

AI Technical Summary

Technical Problem

In the existing metal layer preparation process of integrated circuit chips, the metal layer wire breakage and local residual aluminum dots are prone to occur, resulting in a decrease in the yield of the finished product.

Method used

By training the finished product yield prediction model based on historical data, identifying the combination of key process parameters and candidate adjustment terms, and optimizing the process parameters of the integrated circuit metal layer lithography process to reduce the occurrence of metal layer wire breaks and local residual aluminum points.

Benefits of technology

The finished product yield of the integrated circuit chip metal layer has been significantly improved, and the occurrence frequency of wire breaks of metal layer and local residual aluminum dots has been reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118800645B_ABST
    Figure CN118800645B_ABST
Patent Text Reader

Abstract

The present application discloses an optimization method and device for integrated circuit metal layer lithography process, belonging to the field of semiconductor process technology. The method trains a finished product yield prediction model through a process parameter set of the integrated circuit metal layer lithography process and historical data of finished product yield, calculates the first degree of influence of each process parameter on the finished product yield to identify at least one key process parameter with a significant first degree of influence on the finished product yield, calculates the second degree of influence of the candidate adjustment item combination of the key process parameters on the finished product yield, and selects the optimal candidate adjustment item combination to optimize the integrated circuit metal layer lithography process. The present application can reduce the problems of metal layer wire breakage, local residual aluminum spots, etc. in the integrated circuit metal layer lithography process, and improve the finished product yield of the metal layer of the integrated circuit chip.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of semiconductor process technology, and in particular to a method and device for optimizing an integrated circuit metal layer lithography process. Background Art

[0002] In the preparation process of semiconductor integrated circuit (IC) chips, the preparation of the metal layer of the integrated circuit chip is the process of depositing a conductive metal film on the integrated circuit chip by chemical or physical treatment methods. At present, the preparation of the metal layer of the integrated circuit chip is usually carried out by vacuum evaporation of metal materials such as aluminum (referred to as aluminum evaporation process) and sputtering to deposit a metal film on the surface of a silicon or silicon dioxide substrate, and then the formation of the metal layer conductive pattern is completed through metal layer photolithography processes such as coating, pre-baking, exposure, development, hardening, corrosion, and degumming. However, this metal layer photolithography process combined with the aluminum evaporation process is prone to problems such as broken metal layer wires and local residual aluminum spots, which reduces the finished product yield of the metal layer of the integrated circuit chip. Therefore, an improved technical solution is urgently needed to solve the above problems. Summary of the invention

[0003] In view of this, the present application proposes an optimization method and device for the integrated circuit metal layer lithography process, which can significantly reduce the problems of metal layer wire breakage, local residual aluminum spots, etc. during the integrated circuit metal layer lithography process, and improve the finished product yield of the integrated circuit chip metal layer.

[0004] In a first aspect, an embodiment of the present application provides a method for optimizing an integrated circuit metal layer lithography process, the method comprising:

[0005] Training a finished product yield prediction model based on a process parameter set of the integrated circuit metal layer lithography process and historical data of finished product yield;

[0006] Based on the finished product yield prediction model, calculating a first influence degree of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield, and identifying at least one key process parameter having a significant first influence degree on the finished product yield from the process parameter set;

[0007] Based on the finished product yield prediction model, the second degree of influence of the candidate adjustment item combination of at least one key process parameter on the finished product yield is calculated, and the optimal candidate adjustment item combination is selected to optimize the integrated circuit metal layer lithography process according to the second degree of influence of the candidate adjustment item combination of at least one key process parameter on the finished product yield.

[0008] In some embodiments, the training of the finished product yield prediction model based on the process parameter set of the integrated circuit metal layer lithography process and the historical data of finished product yield comprises:

[0009] The process parameter set of the integrated circuit metal layer lithography process and the first historical data of the finished product yield are used as a training set to train a gradient enhancement decision tree model as the finished product yield prediction model.

[0010] In some embodiments, the step of calculating a first degree of influence of each process parameter in a process parameter set of the integrated circuit metal layer lithography process on the finished product yield based on the finished product yield prediction model, and identifying at least one key process parameter having a significant first degree of influence on the finished product yield from the process parameter set includes:

[0011] Based on the finished product yield prediction model, a first average influence characteristic value of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield is calculated, and the first average influence characteristic value of each process parameter on the finished product yield is used to characterize the first influence degree of each process parameter on the finished product yield;

[0012] According to the magnitude of the first average influence characteristic value of each process parameter on the finished product yield, at least one key process parameter having a top first average influence characteristic value on the finished product yield is identified from the process parameter set of the integrated circuit metal layer lithography process.

[0013] In some embodiments, the calculating, based on the finished product yield prediction model, a first average impact characteristic value of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield includes:

[0014] Randomly select a group of batch sample values ​​of each process parameter in the process parameter set of the integrated circuit metal layer lithography process, and calculate a first SHAP value of each batch sample value in the group of batch sample values ​​of each process parameter to the finished product yield based on the finished product yield prediction model;

[0015] The absolute average of the first SHAP value of each batch of sample values ​​in a group of batch sample values ​​of each process parameter on the finished product yield is calculated to obtain a first average influence characteristic value of each process parameter on the finished product yield.

[0016] In some embodiments, the step of identifying at least one key process parameter having a top first average influence characteristic value on the finished product yield from a set of process parameters of the integrated circuit metal layer lithography process according to the magnitude of the first average influence characteristic value of each process parameter on the finished product yield comprises:

[0017] Calculate the cumulative measurement value of the first average impact characteristic value of the process parameters included in each process of the integrated circuit metal layer lithography process on the finished product yield;

[0018] A predetermined number of target process steps having the highest cumulative measurement values ​​for the first average influencing characteristic value on the finished product yield are selected, and process parameters included in the target process steps are used as the at least one key process parameter.

[0019] In some embodiments, the calculating, based on the finished product yield prediction model, a second degree of influence of the candidate adjustment item combination of the at least one key process parameter on the finished product yield, and selecting the optimal candidate adjustment item combination to optimize the integrated circuit metal layer lithography process according to the second degree of influence of the candidate adjustment item combination of the at least one key process parameter on the finished product yield, includes:

[0020] Based on the finished product yield prediction model, calculating a second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield, the second average impact characteristic value of the candidate adjustment item combination on the finished product yield being used to characterize a second degree of impact of the candidate adjustment item combination on the finished product yield;

[0021] A candidate adjustment item combination with a maximum second average influencing characteristic value of the at least one key process parameter is selected to optimize the integrated circuit metal layer lithography process.

[0022] In some embodiments, the calculating, based on the finished product yield prediction model, a second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield includes:

[0023] Selecting a group of batch sample values ​​including any candidate adjustment item combination of the at least one key process parameter, and calculating a second SHAP value of the candidate adjustment item of each key process parameter in the at least one key process parameter for the finished product yield based on the finished product yield prediction model;

[0024] The average cumulative value of the second SHAP value of the finished product yield rate of the candidate adjustment item combination of the at least one key process parameter is calculated to obtain the second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield rate.

[0025] In some embodiments, the training of the product yield prediction model based on the process parameter set of the integrated circuit metal layer lithography process and the historical data of product yield further includes:

[0026] The process parameter set of the integrated circuit metal layer lithography process and the second historical data of the finished product yield are used as a test set to verify the gradient enhancement decision tree model to obtain the optimal finished product yield prediction model.

[0027] In some embodiments, the process parameter set of the integrated circuit metal layer lithography process includes coating speed, coating time, pre-baking temperature, pre-baking time, exposure time, development time, hardening temperature, hardening time, etching temperature, etching time, and stripping time.

[0028] In a second aspect, the present application also provides an optimization device for an integrated circuit metal layer lithography process, the device comprising:

[0029] A model training unit, used for training a finished product yield prediction model based on a process parameter set of the integrated circuit metal layer lithography process and historical data of finished product yield;

[0030] a parameter acquisition unit, configured to calculate, based on the finished product yield prediction model, a first degree of influence of each process parameter in a process parameter set of the integrated circuit metal layer lithography process on the finished product yield, and identify at least one key process parameter having a significant first degree of influence on the finished product yield from the process parameter set;

[0031] A process optimization unit is used to calculate the second degree of influence of the candidate adjustment item combination of at least one key process parameter on the finished product yield based on the finished product yield prediction model, and select the optimal candidate adjustment item combination to optimize the integrated circuit metal layer lithography process according to the second degree of influence of the candidate adjustment item combination of at least one key process parameter on the finished product yield.

[0032] This application can at least achieve the following beneficial effects:

[0033] In the embodiment of the present application, a finished product yield prediction model is trained based on a process parameter set of an integrated circuit metal layer lithography process and historical data of finished product yield, and a first degree of influence of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield is calculated based on the finished product yield prediction model, and at least one key process parameter having a significant first degree of influence on the finished product yield is identified from the process parameter set; subsequently, based on the finished product yield prediction model, a second degree of influence of a candidate adjustment item combination of the at least one key process parameter on the finished product yield is calculated, and an optimal candidate adjustment item combination is selected according to the second degree of influence of the candidate adjustment item combination of the at least one key process parameter on the finished product yield to optimize the integrated circuit metal layer lithography process, so that the key process and key process parameters having a significant influence on the metal layer finished product yield of the integrated circuit chip can be identified based on machine learning of the historical data of the integrated circuit metal layer lithography process, and the candidate adjustment item combination of the key process parameters having the greatest positive degree of influence on the finished product yield is identified to optimize the lithography process, which can significantly reduce the metal layer wire breakage, local residual aluminum spots and other undesirable problems in the integrated circuit metal layer lithography process, and improve the finished product yield of the metal layer of the integrated circuit chip. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments of the present application. It should be understood that the following drawings only illustrate certain embodiments of the present application and should not be regarded as limiting the scope of the present application.

[0035] Figure 1 It is a schematic flow chart of a method for optimizing an integrated circuit metal layer photolithography process according to the first embodiment of the present application;

[0036] Figure 2 It is a partial flow diagram of a method for optimizing an integrated circuit metal layer photolithography process according to the second embodiment of the present application;

[0037] Figure 3 It is a partial flow diagram of a method for optimizing an integrated circuit metal layer photolithography process according to the third embodiment of the present application;

[0038] Figure 4 is a partial flow chart of an optimization method for an integrated circuit metal layer photolithography process according to a fourth embodiment of the present application;

[0039] Figure 5 is a partial flow chart of an optimization method for an integrated circuit metal layer photolithography process according to a fifth embodiment of the present application;

[0040] Figure 6is an exemplary flow chart of the integrated circuit metal layer photolithography process described in the embodiment of the present application;

[0041] Figure 7 is a schematic structural diagram of an optimization device for an integrated circuit metal layer photolithography process according to the first embodiment of the present application;

[0042] Figure 8 is a partial structural schematic diagram of an integrated circuit metal layer photolithography process optimization device according to the second embodiment of the present application;

[0043] Fig. 9 is a partial structural schematic diagram of an integrated circuit metal layer photolithography process optimization device according to the third embodiment of the present application;

[0044] Fig.10 is a partial structural schematic diagram of an integrated circuit metal layer photolithography process optimization device according to a fourth embodiment of the present application;

[0045] Fig.11 It is a partial structural schematic diagram of an optimization device for an integrated circuit metal layer lithography process according to the fifth embodiment of the present application. DETAILED DESCRIPTION

[0046] To make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present application. However, it should be understood that the described embodiments are only some exemplary embodiments of the present application, rather than all embodiments, and therefore the following detailed description of the embodiments of the present application is not intended to limit the scope of protection claimed in the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0047] It should be noted that the terms "first", "second", etc. in the specification and claims of this application are only used to distinguish and describe similar objects, rather than to describe a specific order or sequence, and cannot be understood as indicating or implying relative importance.

[0048] As mentioned above, the metal layer of the integrated circuit chip is usually prepared by vacuum evaporation of metal materials such as aluminum (referred to as aluminum evaporation process) and sputtering to deposit a metal film on the surface of a silicon or silicon dioxide substrate, and then the formation of the metal layer conductive pattern is completed through metal layer photolithography processes such as coating, pre-baking, exposure, development, hardening, etching, and degumming. However, this metal layer photolithography process combined with the aluminum evaporation process is prone to undesirable problems such as broken metal layer wires and local residual aluminum spots, which reduces the finished product yield of the metal layer of the integrated circuit chip. To this end, the embodiment of the present application proposes an optimization method and device for the metal layer photolithography process of an integrated circuit, which can significantly reduce the undesirable problems such as broken metal layer wires and local residual aluminum spots during the metal layer photolithography process of the integrated circuit, and improve the finished product yield of the metal layer of the integrated circuit chip.

[0049] Figure 1 FIG. 1 is a flow chart of an optimization method for an integrated circuit metal layer photolithography process according to the first embodiment of the present application. Figure 1 As shown, the method comprises the following steps:

[0050] Step S110 , training a finished product yield prediction model based on the process parameter set of the integrated circuit metal layer lithography process and historical data of finished product yield.

[0051] In the first embodiment, the integrated circuit metal layer lithography process may include multiple processes such as coating, pre-baking, exposure, development, hardening, etching, and debonding. Each process includes at least one process parameter, and the process parameters included in each process constitute a process parameter set. Assuming that the integrated circuit metal layer lithography process includes N process processes, the process parameter set included in all process processes can be expressed as ,in represents the m process parameters in the process parameter set, Indicates the first i process parameters, m, m is an integer greater than 1.

[0052] After testing, the corresponding finished product yield rate can be obtained for each batch of finished products generated by the integrated circuit metal layer photolithography process. The finished product yield rate represents the ratio of the number of qualified finished products in each batch to the total number of finished products. i Process parameters The sample value of the zth batch is expressed as , then the zth batch sample set of the process parameter set can be expressed as , the finished product yield of the zth batch is expressed as , a training set is constructed based on the process parameter sets and historical data of finished product yield of multiple batches, and the historical data set of process parameter sets is used As the input feature vector, the corresponding finished product yield data set To output label values, a finished product yield prediction model can be trained.

[0053] In some embodiments, considering the nonlinear correlation characteristics between the process parameters and the finished product yield in the integrated circuit metal layer lithography process, the finished product yield prediction model can be obtained based on the training of the gradient boosted decision tree model (GBDT). As one of the examples, the gradient boosted decision tree model can be implemented based on the LightGBM model. The LightGBM model is a specific algorithm implementation of the gradient boosted decision tree model, which supports efficient parallel training and has the advantages of faster training speed, lower memory consumption, better accuracy, support for distributed processing of massive data, etc. The LightGBM model uses the histogram algorithm to find the best splitting node of the decision tree, sorts each dimensional feature in the training sample of the process parameter set, and divides each dimensional feature into histograms to construct a decision tree.

[0054] Furthermore, a test set is constructed based on the process parameter sets of multiple batches and the historical data of the finished product yield, and the performance of the LightGBM model is verified based on the loss function to achieve the expected performance indicators. An exemplary loss function can be expressed as: ,in represents the predicted value of the finished product yield of the zth batch of samples, Represents the root mean square error loss function of a single decision tree of the LightGBM model.

[0055] Step S120, based on the finished product yield prediction model, calculate the first influence degree of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield, and identify at least one key process parameter that has a significant first influence degree on the finished product yield from the process parameter set.

[0056] In the first embodiment, after training the finished product yield prediction model based on the process parameter set of the integrated circuit metal layer lithography process and the historical data of the finished product yield, it is necessary to further calculate the first degree of influence of each process parameter in the process parameter set on the finished product yield based on the finished product yield prediction model. Since the degree of influence of each process parameter in the process parameter set on the finished product yield of the metal layer preparation is different, in practice, it is often necessary to rely on the experiments and experience of technicians to determine which process parameters in the process parameter set have a greater degree of influence on the finished product yield of the metal layer preparation. This method is extremely low for the optimization efficiency of the integrated circuit metal layer lithography process and is also very wasteful of production resources. In order to solve this problem, the interpretable model algorithm of the gradient enhancement decision tree model can be combined in this embodiment to realize the measurement of the degree of influence of each process parameter in the process parameter set on the finished product yield.

[0057] Specifically, in the second embodiment, if Figure 2 As shown, in step S120, based on the finished product yield prediction model, calculating the first influence degree of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield, and identifying at least one key process parameter having a significant influence degree on the finished product yield from the process parameter set, may include:

[0058] Step S210, based on the finished product yield prediction model, calculating a first average influence characteristic value of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield, wherein the first average influence characteristic value of each process parameter on the finished product yield is used to characterize a first influence degree of each process parameter on the finished product yield;

[0059] Step S220, based on the size of the first average impact characteristic value of each process parameter on the finished product yield, identify at least one key process parameter having a top first average impact characteristic value on the finished product yield from the process parameter set of the integrated circuit metal layer lithography process.

[0060] In some embodiments, the first average influence characteristic value of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield can be characterized by the average SHAP value of a set of random sample values ​​of each process parameter. The SHAP value is an interpretable model algorithm developed based on Python for explaining the degree of influence of feature variables on predicted values ​​in machine learning models, and can be used to measure the degree of influence of each input feature, i.e., process parameter, on the finished product yield in the finished product yield prediction model.

[0061] In the third embodiment, if Figure 3As shown, in step S210 of this embodiment, based on the finished product yield prediction model, calculating the first average influence characteristic value of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield may include:

[0062] Step S310, randomly selecting a group of batch sample values ​​of each process parameter in the process parameter set of the integrated circuit metal layer lithography process, and calculating a first SHAP value of each batch sample value in the group of batch sample values ​​of each process parameter to the finished product yield based on the finished product yield prediction model;

[0063] Step S320, calculating the absolute average of the first SHAP value of each batch of sample values ​​in a group of batch sample values ​​of each process parameter on the finished product yield, to obtain the first average impact characteristic value of each process parameter on the finished product yield.

[0064] In the third embodiment, first, each process parameter in the process parameter set of the integrated circuit metal layer lithography process is randomly selected. A set of batch sample values , Q is each process parameter The number of sample values ​​in a batch. Each process parameter Each batch of sample values ​​in a set of batch sample values The first SHAP value of the finished product yield can be calculated by the following formula (1):

[0065] (1).

[0066] in, Represents each process parameter Each batch of sample values ​​in a set of batch sample values a first SHAP value for the finished product yield; represents the j-th batch sample set of the process parameter set; S represents the j-th batch sample set of the process parameter set Each process parameter The j-th batch sample value A subset of sequentially related alliances that satisfies ; m represents the number of process parameters in the process parameter set; Indicates that S and The predicted value of the finished product yield prediction model when the union of is the input feature vector; It represents the predicted value of the finished product yield prediction model when S is used as the input feature vector.

[0067] Since the first SHAP value calculated by the above formula may include positive or negative values, which respectively represent the positive or negative influence of the sample value of each process parameter on the predicted value of the finished product yield, in order to avoid the random error of the first SHAP value of a single sample of each process parameter, in the embodiment of the present application, each process parameter is calculated and obtained. Each batch of sample values ​​in a set of batch sample values Based on the first SHAP value of the finished product yield, each process parameter is further calculated Each batch of sample values ​​in a set of batch sample values The average absolute value of the first SHAP value of the finished product yield is used as the average absolute value of the first SHAP value of the finished product yield. The first average impact characteristic value on the finished product yield. Each of the process parameters The calculation formula (2) of the first average influencing characteristic value on the finished product yield is as follows:

[0068] (2).

[0069] It can be seen that each process parameter calculated based on the above formula The first average impact characteristic value on the finished product yield can be used to characterize the importance of the impact of each process parameter on the finished product yield, and the importance of the impact is independent of the positive or negative impact of each process parameter on the finished product yield.

[0070] Furthermore, this embodiment can be based on each process parameter The magnitude of the first average influencing characteristic value of the finished product yield is determined by identifying at least one key process parameter having a top first average influencing characteristic value on the finished product yield from a process parameter set of the integrated circuit metal layer lithography process.

[0071] In the fourth embodiment, if Figure 4 As shown, in step S220 of this embodiment, according to the magnitude of the first average influence characteristic value of each process parameter on the finished product yield, identifying at least one key process parameter with a top first average influence characteristic value on the finished product yield from the process parameter set of the integrated circuit metal layer lithography process may include:

[0072] Step S410, calculating a cumulative measurement value of a first average impact characteristic value of a process parameter included in each process of the integrated circuit metal layer lithography process on the finished product yield;

[0073] Step S420 , selecting a predetermined number of target process steps having the top cumulative measurement values ​​of the first average influencing characteristic value on the finished product yield, and using process parameters included in the target process steps as the at least one key process parameter.

[0074] In a fourth embodiment, when identifying at least one key process parameter whose first average influence characteristic value on the finished product yield is at the top from the process parameter set of the integrated circuit metal layer lithography process, a predetermined number of target process processes in the integrated circuit metal layer lithography process that have the most significant impact on the finished product yield can be identified based on the cumulative measurement value of the first average influence characteristic value of the process parameters included in each process process on the finished product yield, and the process parameters included in the target process process can be used as the at least one key process parameter.

[0075] Assume that the index set of the process parameter subset included in the kth process among the N process processes of the integrated circuit metal layer lithography process is expressed as , ; then the cumulative measurement value of the first average impact characteristic value of the process parameters contained in each process on the finished product yield It can be calculated by the following formula (3):

[0076] (3).

[0077] In the fourth embodiment, the cumulative measurement value of the first average impact characteristic value of the process parameters included in each process on the finished product yield The degree of influence of each process on the finished product yield as a whole is characterized, and a predetermined number of target process processes with the greatest influence on the finished product yield in the integrated circuit metal layer lithography process can be identified.

[0078] In the fourth embodiment, the predetermined number is an empirical value set based on the number of all process steps in the integrated circuit metal layer lithography process. Assuming the predetermined number is 3, it indicates that the top three target process steps with the highest cumulative measurement values ​​of the first average influencing characteristic value on the finished product yield in the integrated circuit metal layer lithography process need to be identified, and the process parameters included in these target process steps are the key process parameters for performing subsequent process optimization.

[0079] Step S130, based on the finished product yield prediction model, calculate the second degree of influence of the candidate adjustment item combination of the at least one key process parameter on the finished product yield, and select the optimal candidate adjustment item combination to optimize the integrated circuit metal layer lithography process according to the second degree of influence of the candidate adjustment item combination of the at least one key process parameter on the finished product yield.

[0080] After obtaining at least one key process parameter with a first average influencing characteristic value on the finished product yield in the aforementioned steps, the second degree of influence of each candidate adjustment item combination of the at least one key process parameter on the finished product yield can be further calculated based on the finished product yield prediction model. It can be understood that the second degree of influence of each candidate adjustment item combination of the at least one key process parameter on the finished product yield directly determines the key action of optimizing the integrated circuit metal layer lithography process. In other words, according to the second degree of influence of each candidate adjustment item combination of the at least one key process parameter on the finished product yield, the candidate adjustment item combination of the at least one key process parameter is selected to optimize the integrated circuit metal layer lithography process, and the finished product yield of the metal layer produced and prepared based on the optimized integrated circuit metal layer lithography process should theoretically be optimal.

[0081] In the fifth embodiment, if Figure 5 As shown, in the step S130, the second degree of influence of the candidate adjustment item combination of the at least one key process parameter on the finished product yield is calculated based on the finished product yield prediction model, and the optimal candidate adjustment item combination is selected according to the second degree of influence of the candidate adjustment item combination of the at least one key process parameter on the finished product yield to optimize the integrated circuit metal layer lithography process, which may include:

[0082] Step S510, based on the finished product yield prediction model, calculating a second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield, wherein the second average impact characteristic value of the candidate adjustment item combination on the finished product yield is used to characterize a second impact degree of the candidate adjustment item combination on the finished product yield;

[0083] Step S520: Select a candidate adjustment item combination having a maximum second average impact characteristic value of the at least one key process parameter to optimize the integrated circuit metal layer lithography process.

[0084] In the fifth embodiment, it is assumed that the index set of the at least one key process parameter is expressed as Each of the at least one key process parameter is expressed as ,in ; Each key process parameter Each candidate adjustment item represents , T is each key process parameter The number of candidate adjustment items for each key process parameter The candidate adjustment items are parameter adjustment items that meet the process specifications. These candidate adjustment items can be obtained based on the baseline parameters according to the predetermined increase and decrease gradient. Each key process parameter The candidate adjustment items may constitute a set of candidate adjustment items.

[0085] First, select any candidate adjustment item combination including the at least one key process parameter A set of batch sample values , R represents any candidate adjustment item combination including the at least one key process parameter A set of indices of batch sample values, where Then, based on the finished product yield prediction model, each of the at least one key process parameter is calculated. Candidate adjustment items The second SHAP value of the finished product yield can be calculated by the following formula (4):

[0086] (4).

[0087] in, Indicates each key process parameter Candidate adjustment items a second SHAP value for the finished product yield; Represents the rth batch sample set of the process parameter set; Represents the rth batch sample set of the process parameter set Each key process parameter Candidate adjustment items A subset of sequentially related alliances that satisfies ; m represents the number of process parameters in the process parameter set; Indicates and The predicted value of the finished product yield prediction model when the union of is the input feature vector; Indicates is the predicted value of the finished product yield prediction model when the feature vector is input.

[0088] Since the second SHAP value calculated by the above formula may include positive or negative values, it represents each key process parameter Candidate adjustment items The degree of positive or negative impact on the predicted value of finished product yield.

[0089] Then, in the embodiment of the present application, each of the key process parameters is calculated. Candidate adjustment items Based on the second SHAP value of the finished product yield, further calculating the candidate adjustment item combination of the at least one key process parameter The average cumulative value of the second SHAP value of the finished product yield is used as the second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield. The calculation formula (5) of the second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield is as follows:

[0090] (5).

[0091] in, A second average impact characteristic value representing a combination of candidate adjustment items of the at least one key process parameter on the finished product yield; Indicates that any combination of candidate adjustment items is included The number of batches of sample values ​​for a set of batches.

[0092] It can be seen that the second average impact characteristic value of the candidate adjustment item combination of at least one key process parameter on the finished product yield calculated based on the above formula can be used to characterize the second impact degree of the candidate adjustment item combination on the finished product yield, and the second impact degree is based on the candidate adjustment item combination of at least one key process parameter. The average cumulative value of the second SHAP value of the finished product yield is calculated, and therefore is related to the degree of positive or negative impact of the candidate adjustment item combination on the finished product yield. Since the degree of positive impact of the candidate adjustment item combination on the finished product yield represents a positive promotion effect on the finished product yield. Therefore, this embodiment can select the candidate adjustment item combination with the largest second average impact characteristic value of the at least one key process parameter from the candidate adjustment item combination with a positive second average impact characteristic value to optimize the integrated circuit metal layer lithography process.

[0093] Combine the following Figure 6 The exemplary specific implementation of the method for optimizing the integrated circuit metal layer photolithography process described in the embodiment of the present application is specifically described. Figure 6 As shown, in this embodiment, the integrated circuit metal layer photolithography process may include multiple processes such as coating, pre-baking, exposure, development, hardening, etching, and stripping. Figure 6 As shown in P1~P7. The above multiple process components include process parameter sets , where the process of gluing P1 includes the process parameters of gluing speed Gluing time ; Pre-bake P2 includes process parameters pre-bake temperature , Pre-baking time ; Exposure P3 includes process parameters exposure time ; Development P4 includes process parameters development time ; Hardening P5 includes process parameters hardening temperature , Hardening time ; Corrosion P6 includes process parameters corrosion temperature , corrosion time ; Degumming P7 includes process parameters degumming time .

[0094] First, a training set is constructed based on the historical data of process parameter sets and finished product yields of multiple batches. The historical data set of the process parameter set is used as the input feature vector, and the corresponding finished product yield data set is used as the output label value. The finished product yield prediction model based on the LightGBM model can be trained.

[0095] Then, randomly select each process parameter in the process parameter set A set of batch sample values, according to the above formulas (1) and (2) to calculate each process parameter The first average impact characteristic value on the finished product yield , the exemplary calculation results arranged in descending order of size are shown in Table 1 below:

[0096] Table 1

[0097]

[0098] Each process parameter calculated above The first average impact characteristic value on the finished product yield Each process parameter Each batch of sample values ​​in a set of batch sample values The absolute average of the first SHAP value of the finished product yield is used to characterize the importance of the impact of each process parameter on the finished product yield, and the importance of the impact is independent of the positive or negative impact of each process parameter on the finished product yield. The first average impact characteristic value on the finished product yield , according to the above formula (3), calculate the cumulative measurement value of the first average impact characteristic value of the process parameters included in each process on the finished product yield , the exemplary calculation results arranged in descending order of size are shown in Table 2 below:

[0099] Table 2

[0100]

[0101] From the above table, we can identify the predetermined number of target processes with the highest cumulative measurement values ​​of the first average influencing characteristic value on the finished product yield, and the process parameters contained in these target processes are the key process parameters for performing subsequent process optimization. For example, when the predetermined number is 3, the three target processes with the highest cumulative measurement values ​​can be identified as coating P1, exposure P3, and development P4, and the key process parameter for performing subsequent process optimization can be obtained as coating speed. Gluing time , exposure time , Development time .

[0102] Further, based on the finished product yield prediction model, the second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield is calculated according to the above formulas (4) and (5): The second average impact characteristic value is calculated based on the average cumulative value of the second SHAP value of the candidate adjustment item combination of at least one key process parameter on the finished product yield, and is therefore related to the degree of positive or negative impact of the candidate adjustment item combination on the finished product yield. Exemplary calculation results, arranged in descending order of size, are shown in Table 3 below:

[0103] Table 3

[0104]

[0105] The table above shows the key process parameters of glue coating speed. Gluing time , exposure time , Development time The second average impact characteristic value of the combination of 3 candidate adjustment items is positive, where the candidate adjustment item combination { , , , The second average impact eigenvalue of Maximum, so that this embodiment can select the candidate adjustment item combination with the largest second average impact characteristic value from the candidate adjustment item combinations with positive second average impact characteristic values ​​{ , , , }Optimize the metal layer photolithography process of the integrated circuit.

[0106] Figure 7FIG. 1 is a schematic diagram of a structure of an optimization device for an integrated circuit metal layer photolithography process according to the first embodiment of the present application. Figure 7 As shown, in a first embodiment, the device comprises the following units:

[0107] A model training unit 710, configured to train a finished product yield prediction model based on a process parameter set of the integrated circuit metal layer lithography process and historical data of finished product yield;

[0108] A parameter acquisition unit 720 is configured to calculate a first degree of influence of each process parameter in a process parameter set of the integrated circuit metal layer lithography process on the finished product yield based on the finished product yield prediction model, and identify at least one key process parameter having a significant first degree of influence on the finished product yield from the process parameter set;

[0109] The process optimization unit 730 is used to calculate the second degree of influence of the candidate adjustment item combination of at least one key process parameter on the finished product yield based on the finished product yield prediction model, and select the optimal candidate adjustment item combination to optimize the integrated circuit metal layer lithography process according to the second degree of influence of the candidate adjustment item combination of at least one key process parameter on the finished product yield.

[0110] In the second embodiment, if Figure 8 As shown, the parameter acquisition unit 720 may include:

[0111] A first eigenvalue calculation unit 810 is used to calculate a first average influence eigenvalue of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield based on the finished product yield prediction model, wherein the first average influence eigenvalue of each process parameter on the finished product yield is used to characterize a first influence degree of each process parameter on the finished product yield;

[0112] The process parameter identification unit 820 is used to identify at least one key process parameter having a first average influence characteristic value on the finished product yield from the process parameter set of the integrated circuit metal layer lithography process according to the size of the first average influence characteristic value of each process parameter on the finished product yield.

[0113] In the third embodiment, if Fig. 9 As shown, the first eigenvalue calculation unit 810 may include:

[0114] A first SHAP value calculation unit 910 is used to randomly select a group of batch sample values ​​of each process parameter in the process parameter set of the integrated circuit metal layer lithography process, and calculate a first SHAP value of each batch sample value in the group of batch sample values ​​of each process parameter for the finished product yield based on the finished product yield prediction model;

[0115] The influence characteristic value calculation unit 920 is used to calculate the average of the absolute values ​​of the first SHAP values ​​of each batch of sample values ​​in a group of batch sample values ​​of each process parameter on the finished product yield, so as to obtain the first average influence characteristic value of each process parameter on the finished product yield.

[0116] In the fourth embodiment, if Fig.10 As shown, the process parameter identification unit 820 may include:

[0117] The cumulative metric calculation unit 1010 is used to calculate the cumulative metric value of the first average influence characteristic value of the process parameters included in each process of the integrated circuit metal layer lithography process on the finished product yield;

[0118] The process identification unit 1020 is used to select a predetermined number of target process steps having the top cumulative measurement values ​​of the first average influencing characteristic value on the finished product yield, and use the process parameters included in the target process steps as the at least one key process parameter.

[0119] In the fifth embodiment, if Fig.11 As shown, the process optimization unit 730 may include:

[0120] A second characteristic value calculation unit 1110 is used to calculate a second average influence characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield based on the finished product yield prediction model, wherein the second average influence characteristic value of the candidate adjustment item combination on the finished product yield is used to characterize a second influence degree of the candidate adjustment item combination on the finished product yield;

[0121] The parameter adjustment item selection unit 1120 is used to select a candidate adjustment item combination with a maximum second average impact characteristic value of the at least one key process parameter to optimize the integrated circuit metal layer lithography process.

[0122] In summary, an optimization method and device for an integrated circuit metal layer lithography process in an embodiment of the present application trains a finished product yield prediction model based on a process parameter set of the integrated circuit metal layer lithography process and historical data on finished product yield, calculates a first degree of influence of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield based on the finished product yield prediction model, and identifies at least one key process parameter that has a significant first degree of influence on the finished product yield from the process parameter set; then, based on the finished product yield prediction model, calculates a second degree of influence of a combination of candidate adjustment items for the at least one key process parameter on the finished product yield, and calculates a second degree of influence of a combination of candidate adjustment items for the at least one key process parameter on the finished product yield. The optimal candidate adjustment item combination is selected based on the second degree of influence of at least one candidate adjustment item combination of key process parameters on the finished product yield to optimize the integrated circuit metal layer lithography process, so that the key process and key process parameters that have a significant impact on the metal layer finished product yield of the integrated circuit chip can be identified based on machine learning of historical data of the integrated circuit metal layer lithography process, and the candidate adjustment item combination of key process parameters that has the greatest positive impact on the finished product yield is identified to optimize the lithography process, which can significantly reduce the problems of broken metal layer wires, local residual aluminum spots, etc. in the integrated circuit metal layer lithography process, and improve the finished product yield of the metal layer of the integrated circuit chip.

[0123] It should be noted that those skilled in the art can understand that the different implementation methods described in the method embodiments of the present application and their explanations and technical effects achieved are also applicable to the device embodiments of the present application and will not be repeated here.

[0124] Furthermore, an embodiment of the present application also proposes an electronic device, which may include: a processor and a memory. The memory stores computer program instructions, and the processor may call the computer program instructions in the memory to execute all or part of the steps of the method described in any embodiment of the present application. The computer program instructions in the above-mentioned memory may be implemented in the form of a software functional unit and may be stored in a computer-readable storage medium when sold or used as an independent product.

[0125] Furthermore, the present application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program. When the computer-readable storage medium is connected to a computer device and the computer program is executed by one or more processors of the computer device, it can execute all or part of the steps of the method described in any embodiment of the present application.

[0126] Furthermore, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which can be executed by one or more processors to perform all or part of the steps of the method described in any embodiment of the present application.

[0127] Through the description of the above implementation methods, those skilled in the art can clearly understand that the various implementation methods of the present application can be implemented by means of software or software combined with necessary general hardware platforms, and of course can also be implemented by hardware functions. Based on such an understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The software product is stored in a storage medium, including several instructions to enable a computer device, such as but not limited to a personal computer, a server, or a network device, to perform all or part of the steps of the method described in any implementation method of the present application. The aforementioned storage medium may include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., various media that can store computer program codes.

[0128] The above describes the exemplary embodiments of the present application. It should be understood that the above exemplary embodiments are not restrictive but illustrative, and the protection scope of the present application is not limited thereto. It should be understood that those skilled in the art can modify and vary the embodiments of the present application without departing from the spirit and scope of the present application, and these modifications and variations should be within the protection scope of the present application.

Claims

1. A method for optimizing the photolithography process of an integrated circuit metal layer, characterized in that: The method comprises: Training a finished product yield prediction model based on the process parameter set of the integrated circuit metal layer lithography process and historical data of finished product yield, wherein the finished product yield prediction model is a gradient enhancement decision tree model; Based on the finished product yield prediction model, a first average influence characteristic value of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield is calculated, and the first average influence characteristic value of each process parameter on the finished product yield is used to characterize the first influence degree of each process parameter on the finished product yield, wherein the first average influence characteristic value is independent of the positive or negative influence degree of each process parameter on the finished product yield; according to the size of the first average influence characteristic value of each process parameter on the finished product yield, at least one key process parameter with a front first average influence characteristic value on the finished product yield is identified from the process parameter set of the integrated circuit metal layer lithography process; Based on the finished product yield prediction model, a second average impact characteristic value of the candidate adjustment item combination of at least one key process parameter on the finished product yield is calculated, and the second average impact characteristic value of the candidate adjustment item combination on the finished product yield is used to characterize the second impact degree of the candidate adjustment item combination on the finished product yield, wherein the second average impact characteristic value is related to the positive or negative impact degree of the candidate adjustment item combination on the finished product yield; the candidate adjustment item combination with the largest second average impact characteristic value of the at least one key process parameter is selected to optimize the integrated circuit metal layer lithography process.

2. The method for optimizing the integrated circuit metal layer photolithography process according to claim 1, characterized in that: The training of the finished product yield prediction model based on the process parameter set of the integrated circuit metal layer lithography process and the historical data of finished product yield comprises: The process parameter set of the integrated circuit metal layer lithography process and the first historical data of the finished product yield are used as a training set to train the gradient enhancement decision tree model as the finished product yield prediction model.

3. The method for optimizing the integrated circuit metal layer photolithography process according to claim 2, characterized in that: The step of calculating the first average impact characteristic value of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield based on the finished product yield prediction model includes: Randomly select a group of batch sample values ​​of each process parameter in the process parameter set of the integrated circuit metal layer lithography process, and calculate a first SHAP value of each batch sample value in the group of batch sample values ​​of each process parameter to the finished product yield based on the finished product yield prediction model; The absolute average of the first SHAP value of each batch of sample values ​​in a group of batch sample values ​​of each process parameter on the finished product yield is calculated to obtain a first average influence characteristic value of each process parameter on the finished product yield.

4. The method for optimizing the integrated circuit metal layer photolithography process according to claim 3, characterized in that: The step of identifying at least one key process parameter having a first average influence characteristic value on the finished product yield from a set of process parameters of the integrated circuit metal layer lithography process according to the magnitude of the first average influence characteristic value of each process parameter on the finished product yield, comprises: Calculate the cumulative measurement value of the first average impact characteristic value of the process parameters included in each process of the integrated circuit metal layer lithography process on the finished product yield; A predetermined number of target process steps having the highest cumulative measurement values ​​for the first average influencing characteristic value on the finished product yield are selected, and process parameters included in the target process steps are used as the at least one key process parameter.

5. The method for optimizing the integrated circuit metal layer photolithography process according to claim 4, characterized in that: The calculating, based on the finished product yield prediction model, a second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield includes: Selecting a group of batch sample values ​​including any candidate adjustment item combination of the at least one key process parameter, and calculating a second SHAP value of the candidate adjustment item of each key process parameter in the at least one key process parameter for the finished product yield based on the finished product yield prediction model; The average cumulative value of the second SHAP value of the finished product yield rate of the candidate adjustment item combination of the at least one key process parameter is calculated to obtain the second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield rate.

6. The method for optimizing the integrated circuit metal layer photolithography process according to claim 2, characterized in that: The training of the finished product yield prediction model based on the process parameter set of the integrated circuit metal layer lithography process and the historical data of finished product yield also includes: The process parameter set of the integrated circuit metal layer lithography process and the second historical data of the finished product yield are used as a test set to verify the gradient enhancement decision tree model to obtain the optimal finished product yield prediction model.

7. The method for optimizing the integrated circuit metal layer photolithography process according to claim 1, characterized in that: The process parameter set of the integrated circuit metal layer lithography process includes coating speed, coating time, pre-baking temperature, pre-baking time, exposure time, development time, hardening temperature, hardening time, etching temperature, etching time, and stripping time.

8. An optimization device for integrated circuit metal layer lithography process, characterized in that: The device comprises: A model training unit, used for training a finished product yield prediction model based on a process parameter set of the integrated circuit metal layer lithography process and historical data of finished product yield, wherein the finished product yield prediction model is a gradient enhancement decision tree model; A parameter acquisition unit is used to calculate, based on the finished product yield prediction model, a first average influence characteristic value of each process parameter in the process parameter set of the integrated circuit metal layer lithography process on the finished product yield, wherein the first average influence characteristic value of each process parameter on the finished product yield is used to characterize a first degree of influence of each process parameter on the finished product yield, wherein the first average influence characteristic value is independent of a positive or negative degree of influence of each process parameter on the finished product yield; and according to the magnitude of the first average influence characteristic value of each process parameter on the finished product yield, identify at least one key process parameter with a top first average influence characteristic value on the finished product yield from the process parameter set of the integrated circuit metal layer lithography process; A process optimization unit is used to calculate, based on the finished product yield prediction model, a second average impact characteristic value of the candidate adjustment item combination of the at least one key process parameter on the finished product yield, wherein the second average impact characteristic value of the candidate adjustment item combination on the finished product yield is used to characterize the second degree of influence of the candidate adjustment item combination on the finished product yield, wherein the second average impact characteristic value is related to the degree of positive or negative influence of the candidate adjustment item combination on the finished product yield; and the candidate adjustment item combination with the largest second average impact characteristic value of the at least one key process parameter is selected to optimize the integrated circuit metal layer lithography process.