Method and system for reducing warping change increment of wafer

Abstract

The embodiment of the invention belongs to the field of semiconductors, and discloses a method and system for reducing warping change increment of a wafer. The method comprises the following steps: acquiring real-time pressure, real-time gas flow and real-time temperature of each preset control area in each furnace tube, acquiring property data of a wafer needing to be subjected to thermal process treatment, and inputting the data into a preset warping prediction model to obtain a warping prediction value of the wafer; and if the warping prediction value exceeds a preset warping threshold value, outputting a pressure adjustment value, a gas flow adjustment value and a temperature adjustment value of each preset control area in each furnace tube through a preset warping prediction model, and respectively sending the pressure adjustment value, the gas flow adjustment value and the temperature adjustment value to a pressure control system, a gas flow control valve and a temperature control system of each preset control area in each furnace tube. According to the scheme, the precision of the wafer thermal processing process can be improved, and the condition of warping and deterioration of the wafer is improved, so that the rejection rate is reduced, the material waste is reduced, and the consistency and stability of a production line are realized.

Description

Technical Field

[0001] The present application relates to the field of semiconductor technology, and more particularly to a method and system for reducing the incremental change in wafer warpage. Background Art

[0002] In the semiconductor silicon wafer process, the wafer is the core material for manufacturing chips, and its quality and stability have a decisive impact on the performance of the final product. In order to improve the overall performance and reliability of the wafer, a variety of process steps are applied in its manufacturing process, including back-sealing treatment and polysilicon thin film process.

[0003] Back-sealing is an important technology in semiconductor manufacturing. It prevents impurity ions in the wafer from overflowing from the crystal lattice and forming contamination under high temperature by forming a uniform film on the back of the wafer. This technology can not only improve the cleanliness of the wafer, but also enhance its mechanical strength and provide stable support for subsequent process steps. In practical applications, atmospheric pressure chemical vapor deposition is often used to form a layer of SiO2 film on the back of the wafer to achieve the back-sealing effect. The polysilicon thin film process is also an indispensable part of semiconductor manufacturing. Polysilicon thin film has a high crystal quality, and its crystallinity can reach more than 90%, and its crystal quality and uniformity can be further improved by optimizing the growth process parameters and selecting suitable substrate materials.

[0004] However, the inventors found that there are at least the following technical problems in the related art:

[0005] During the wafer manufacturing process, especially during thermal processing, the polysilicon layer of the wafer may be affected by various factors such as processing equipment, process and external environment, resulting in the deterioration of the entire wafer warpage value. Warpage is an important parameter that describes the change in the shape of the wafer surface. It reflects the degree of bending and distortion of the wafer without external force. When the warpage value exceeds a certain range, the wafer may not meet the requirements of subsequent process steps, and may even cause the entire wafer to be scrapped. Summary of the invention

[0006] One purpose of the present application is to provide a method for reducing the incremental change in wafer warpage, at least to solve the problem that in the prior art, during the manufacturing process of wafers, especially during thermal process treatment, the growth of the polysilicon layer of the wafer may be affected by various factors such as processing equipment, process and external environment, resulting in the deterioration of the entire warpage value of the wafer. Warpage is an important parameter that describes the change in the surface shape of the wafer, which reflects the degree of bending and distortion of the wafer without external force. When the warpage value exceeds a certain range, the wafer may not meet the requirements of subsequent process steps, and may even lead to the scrapping of the entire wafer.

[0007] To achieve the above objectives, some embodiments of the present application provide the following aspects:

[0008] In a first aspect, some embodiments of the present application provide a method for reducing a wafer warpage variation increment, the method comprising:

[0009] Acquire the real-time pressure, real-time gas flow rate and real-time temperature of each preset control area in each furnace tube, and acquire the property data of the wafer to be subjected to the thermal process, and input the real-time pressure, real-time gas flow rate, real-time temperature and property data into the preset warpage prediction model to obtain the warpage prediction value of the wafer to be subjected to the thermal process;

[0010] If the warping prediction value exceeds the preset warping threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube are output through the preset warping prediction model, and the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube.

[0011] According to a second aspect of the present application, a system for reducing the incremental change in wafer warpage is provided, the system comprising:

[0012] A warpage prediction module is used to obtain the real-time pressure, real-time gas flow rate and real-time temperature of each preset control area in each furnace tube, and to obtain the property data of the wafer to be subjected to the thermal process, and input the real-time pressure, real-time gas flow rate, real-time temperature and property data into a preset warpage prediction model to obtain the warpage prediction value of the wafer to be subjected to the thermal process;

[0013] The adjustment value confirmation module is used to output the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube through a preset warping prediction model if the warping prediction value exceeds a preset warping threshold value, and send the pressure adjustment value to the pressure control system of each preset control zone in each furnace tube, send the gas flow adjustment value to the gas flow control valve of each preset control zone in each furnace tube, and send the temperature adjustment value to the temperature control system of each preset control zone in each furnace tube.

[0014] Compared with the related art, in the solution provided in the embodiment of the present application, the real-time pressure, real-time gas flow rate and real-time temperature of each preset control zone in each furnace tube are obtained, and the property data of the wafer that needs to be subjected to the thermal process is obtained, and the real-time pressure, real-time gas flow rate, real-time temperature and property data are input into the preset warping prediction model to obtain the warping prediction value of the wafer that needs to be subjected to the thermal process; if the warping prediction value exceeds the preset warping threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube are output through the preset warping prediction model, and the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube. Through the above-mentioned method of reducing the incremental change of wafer warpage, by realizing this automatic control and adjustment based on real-time data and warpage prediction model, the accuracy of the wafer thermal process can be significantly improved, and the deterioration of wafer warpage can be improved, thereby reducing the scrap rate, reducing material waste, and achieving high consistency and stability of the production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0016] Figure 1 is an exemplary flow chart of a method for reducing the increment of wafer warpage variation provided in Embodiment 1 of the present application;

[0017] Figure 2 is an exemplary flow chart of a method for reducing the increment of wafer warpage variation provided in the second embodiment of the present application;

[0018] Figure 3 is an exemplary flow chart of a method for reducing the increment of wafer warpage variation provided in Embodiment 3 of the present application;

[0019] Figure 4 This is an exemplary structural diagram of a system for reducing the incremental change in wafer warpage provided in Example 4 of the present application. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0021] First embodiment

[0022] The first embodiment of the present application relates to a method for reducing the increment of wafer warpage. Figure 1 As shown, the method may include the following steps:

[0023] S101, obtaining the real-time pressure, real-time gas flow rate and real-time temperature of each preset control area in each furnace tube, and obtaining the property data of the wafer that needs to be subjected to the thermal process, and inputting the real-time pressure, real-time gas flow rate, real-time temperature and property data into a preset warping prediction model to obtain the warping prediction value of the wafer that needs to be subjected to the thermal process.

[0024] First of all, the usage scenario of this solution can be to predict the warpage value of the wafer that needs to undergo thermal process according to the real-time pressure, real-time gas flow and real-time temperature of each preset control area in each furnace tube through the model, and when the predicted value exceeds a certain threshold, the model outputs the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control area in each furnace tube, so as to control the final warpage value of the product within the qualified range.

[0025] Based on the above usage scenarios, it can be understood that the executor of the present application can be a system for reducing the incremental change in wafer warpage, and no excessive limitations are made here.

[0026] A furnace tube can be part of a thermal processing device, typically used for heating and temperature control of wafers. A furnace can contain multiple furnace tubes (4-8 tubes), through which the gas supply, heating, and cooling systems circulate gas in the furnace to ensure uniform temperature and airflow conditions for the wafers during the process. Each furnace tube may be divided into multiple control zones.

[0027] The preset control zones are the partitions in the furnace, which are used to independently control and adjust the temperature, pressure and gas flow of each zone. The division of these control zones is set according to the heat treatment requirements in the furnace. For example, a furnace tube may have multiple zones, such as the air inlet zone, the middle heating zone and the tail cooling zone, and the heat treatment parameters (temperature, gas flow, pressure) in each zone may require different control strategies.

[0028] The real-time pressure can represent the gas pressure in each preset control area in each furnace tube, which is usually monitored and fed back in real time by a pressure sensor.

[0029] The real-time gas flow rate may refer to the gas flow rate in each preset control zone in each furnace tube, which determines the distribution and flow efficiency of the gas in the furnace tube.

[0030] The real-time temperature can represent the real-time temperature in each preset control zone in each furnace tube, which is usually collected by a temperature sensor and used to control the thermal treatment process of the wafer.

[0031] Property data may refer to the material and process characteristics of the wafer during the heat treatment process. Specifically, it may include wafer material characteristics (such as silicon wafer type, thickness, etc.), surface coating information (such as polysilicon film thickness, coating uniformity, etc.), and wafer size (such as diameter, thickness).

[0032] The gas pressure in each preset control zone can be monitored in real time by the pressure sensor installed in the furnace tube. The pressure sensor can capture the gas pressure fluctuations in different furnace tube partitions, helping to determine whether there is too high or too low gas flow pressure, which affects the uniformity of heat treatment. The flow rate of gas in each partition of the furnace tube is measured by flow sensors. These sensors provide flow data based on the gas flow in the furnace tube to ensure that the gas distribution meets the set requirements. The gas flow in different partitions affects the temperature distribution and coating uniformity. Temperature sensors are placed in each control zone to monitor the temperature changes in the furnace. The data collected by these sensors can reflect the real-time fluctuations in the temperature of each partition and help adjust the temperature control system in the furnace to avoid excessively high or low temperatures. Then the property data of the wafer that needs to be processed by the thermal process is read from the database, and then the above real-time data and the property data of the wafer are input into the preset warpage prediction model. The warpage prediction branch of the warpage prediction model predicts the warpage value that may occur during the heat treatment process by training historical data, airflow, temperature, pressure, wafer characteristics and other inputs.

[0033] S102, if the warpage prediction value exceeds the preset warpage threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube are output through the preset warpage prediction model, and the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube.

[0034] The preset warpage threshold can be a defined value that indicates the maximum allowable warpage of the wafer during the thermal process. When the predicted value exceeds this threshold, it indicates that the warpage of the wafer has exceeded the normal range, which may lead to defective products or eventual scrapping. This value is usually set through historical data analysis, empirical formulas or process requirements to ensure the quality of the wafer.

[0035] The pressure adjustment value refers to the gas pressure value that needs to be adjusted when the real-time prediction model detects that the warpage value exceeds the standard. This value is output by the prediction model and used to correct the gas pressure of each preset control area in each furnace tube, thereby optimizing the gas flow distribution and avoiding warpage caused by uneven temperature distribution.

[0036] The gas flow adjustment value may refer to the necessary adjustment of the gas flow in each preset control zone in each furnace tube. The gas flow directly affects the flow of gas in the furnace and the uniformity of temperature, so the adjustment of the flow is one of the key steps in controlling the heat treatment process.

[0037] The temperature adjustment value may refer to the necessary adjustment of the temperature of each preset control zone in each furnace tube based on the warpage prediction result. Temperature fluctuation is an important factor causing wafer warpage. Temperature non-uniformity can lead to inconsistent thermal expansion of the wafer, which in turn causes warpage.

[0038] The pressure control system can be a device and technology used to control the gas pressure in the furnace. It ensures that the gas pressure in each preset control area is maintained within the set range by adjusting the pressure valve, gas flow regulator and other equipment.

[0039] The gas flow control valve is a device that can adjust the gas flow in the furnace as needed. By adjusting the valve, the gas flow is controlled to help achieve uniform distribution of the gas flow in the furnace, thereby avoiding local temperature anomalies.

[0040] The temperature control system can be an automated system that controls the temperature of each zone in the furnace. Through the temperature sensors and heating units of the temperature control system, the system can accurately adjust the temperature in the furnace to ensure that the temperature of each zone meets the process requirements.

[0041] If the warpage prediction value exceeds the preset threshold, the corresponding adjustment value (pressure, gas flow, temperature adjustment value) can be output according to the adjustment value calculation branch of the preset warpage prediction model. Then the output adjustment value is sent to the corresponding control system of each preset control area in the furnace tube through wireless communication technology, where the pressure control system: adjusts the pressure. The gas flow control valve: adjusts the gas flow. The temperature control system: adjusts the temperature.

[0042] It is not difficult to find that compared with the related art, in the solution provided in the embodiment of the present application, the real-time pressure, real-time gas flow rate and real-time temperature of each preset control zone in each furnace tube are obtained, and the property data of the wafer that needs to be subjected to thermal process treatment are obtained, and the real-time pressure, real-time gas flow rate, real-time temperature and property data are input into the preset warping prediction model to obtain the warping prediction value of the wafer that needs to be subjected to thermal process treatment; if the warping prediction value exceeds the preset warping threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube are output through the preset warping prediction model, and the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube. Through the above-mentioned method of reducing the incremental change of wafer warpage, by realizing this automatic control and adjustment based on real-time data and warpage prediction model, the accuracy of the wafer thermal process can be significantly improved, and the deterioration of wafer warpage can be improved, thereby reducing the scrap rate, reducing material waste, and achieving high consistency and stability of the production line.

[0043] On the basis of the above technical solution, optionally, after sending the gas flow adjustment value to the gas flow control valve of each preset control zone in each furnace tube, the method further includes:

[0044] Real-time monitoring to determine whether a preset collection time point has been reached. If the preset collection time point has been reached, the real-time pressure, real-time gas flow rate, and real-time temperature of each preset control area in each furnace tube are re-acquired, and the real-time pressure, real-time gas flow rate, real-time temperature, and property data are re-input into a preset warpage prediction model to update the warpage prediction value.

[0045] If the updated warpage prediction value exceeds the preset warpage threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube are re-output through the preset warpage prediction model, and the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube until the thermal process of the wafer is completed.

[0046] In this solution, the preset collection time point can be a time point during the thermal process when the system needs to collect data and update control parameters regularly or on demand according to predetermined time intervals or conditions.

[0047] It is possible to monitor in real time whether the current time has reached the preset collection time point. Specifically, a timer can be used to track time, and when the time reaches the preset collection period, data collection is triggered. Then the data is collected again and input into the model to obtain the warpage prediction value again. If it still exceeds the preset warpage threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value are re-output through the preset warpage prediction model, and re-sent to the pressure control system, gas flow control valve and temperature control system for readjustment, and then it will continue to monitor and perform periodic data collection and adjustment operations according to the set collection time point until the wafer completes the thermal process. Each time the collection time point is reached, the above steps will be re-executed to ensure that the wafer is always in the best thermal treatment state until the thermal process of the wafer is completed.

[0048] Based on the above technical solution, the optional, preset warpage prediction model training process includes:

[0049] Obtaining historical thermal process processing records, and determining historical pressure, historical gas flow, historical temperature, historical property data of wafers to be subjected to thermal process processing, and a first historical warpage value of each preset control zone in each furnace tube according to the historical thermal process processing records;

[0050] Determine a first historical warpage value exceeding a preset warpage threshold as a second historical warpage value, and obtain a historical pressure adjustment value, a historical gas flow adjustment value, and a historical temperature adjustment value of each preset control zone in each furnace tube corresponding to the second historical warpage value;

[0051] Creating a first data set according to the historical pressure, historical gas flow, historical temperature of each preset control zone in each furnace tube and the historical property data of the wafer to be subjected to the thermal process, and labeling the first data set with a warpage value label according to the first historical warpage value;

[0052] Creating a second data set according to the historical pressure, historical gas flow, historical temperature of each preset control zone in each furnace tube corresponding to the second historical warpage value and the historical property data of the wafer to be subjected to the thermal process, marking the pressure label of the second data set according to the historical pressure, marking the gas flow label of the second data set according to the historical gas flow, and marking the temperature label of the second data set according to the historical temperature;

[0053] Constructing a warpage prediction model, and training a warpage prediction branch of the warpage prediction model according to the first data set and the warpage value label;

[0054] The adjustment value calculation branch of the warpage prediction model is trained according to the second data set, the pressure label, the gas flow label, and the temperature label.

[0055] In this solution, the historical thermal process record may refer to the detailed data of each thermal process of the wafer, including the furnace state (pressure, gas flow, temperature, etc.), wafer properties, processing parameters, and warpage value during each thermal process. These data are usually derived from previous production batches.

[0056] The historical pressure, historical gas flow, and historical temperature may be the working conditions (such as pressure, gas flow, and temperature) of each preset control zone in each furnace tube during the past thermal process, which have a direct impact on the thermal treatment effect and warping of the wafer.

[0057] The historical property data may be property information related to the wafer during thermal process treatment.

[0058] The first historical warpage value may refer to a warpage value of the wafer during the thermal treatment process.

[0059] The second historical warp value may be the first historical warp value that exceeds a preset warp threshold.

[0060] The historical pressure adjustment value, the historical gas flow adjustment value, and the historical temperature adjustment value can be the adjustment values ​​made for excessive warpage in the historical records, with the purpose of correcting the undesirable environmental conditions. They are the results of the warpage value adjustment obtained after the previous thermal process treatment.

[0061] The first data set may include all data from the historical thermal process, including pressure, gas flow, temperature, and wafer property data of each preset control zone in each furnace tube, as well as the warpage value of the batch. This data set is used to train the warpage prediction branch of the warpage prediction model.

[0062] The warpage value label may refer to the target output value of each data point in the first data set, that is, the actual warpage value of the wafer during the thermal treatment process. It is the target variable in supervised learning.

[0063] The second data set may be a data set related to the second historical warpage value, and includes pressure, gas flow, temperature of each preset control zone in each furnace tube in the furnace body, and property data of the wafer.

[0064] The pressure label, gas flow label, and temperature label may represent the adjusted pressure value, gas flow value, and temperature value corresponding to each sample in the second data set. These labels are used for the adjustment value calculation branch of the training model.

[0065] The warpage prediction branch may refer to a submodule in the warpage prediction model, which uses data such as pressure, gas flow, temperature, etc. of each preset control zone in each furnace tube and property data of the wafer to predict the warpage value.

[0066] The adjustment value calculation branch may refer to another part in the warpage prediction model, the purpose of which is to calculate corresponding adjustment values ​​of pressure, gas flow rate and temperature according to the predicted warpage value to correct the warpage problem.

[0067] Historical thermal process records can be obtained, which contain the historical pressure, gas flow, temperature of each preset control zone in the furnace tube, as well as the historical property data and warpage value of the wafer. By analyzing the historical data, the first historical warpage value that exceeds the preset warpage threshold is determined, and it is used as the second historical warpage value to obtain the corresponding pressure, gas flow and temperature adjustment values. Then, based on these historical data, two data sets are created: the first data set contains historical environmental data and wafer property data, and the first historical warpage value is marked as a label; the second data set is marked with historical pressure, gas flow and temperature adjustment values ​​as labels. Then, the first data set and the warpage value label are used to train the warpage prediction branch, and the second data set and the adjustment value label are used to train the adjustment value calculation branch, and finally a comprehensive warpage prediction model is constructed to accurately predict the warpage of the wafer in the thermal process and make necessary adjustments.

[0068] Second embodiment

[0069] The second embodiment of the present application relates to a method for reducing the increment of wafer warpage change. The second embodiment is an improvement on the first embodiment. The second embodiment is substantially the same as the first embodiment, and the main difference is that it specifically includes the following steps:

[0070] S201, obtaining the real-time pressure, real-time gas flow rate and real-time temperature of each preset control area in each furnace tube, and obtaining the property data of the wafer that needs to be subjected to the thermal process, and inputting the real-time pressure, real-time gas flow rate, real-time temperature and property data into a preset warping prediction model to obtain the warping prediction value of the wafer that needs to be subjected to the thermal process.

[0071] S202, if the warpage prediction value exceeds the preset warpage threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube are output through the preset warpage prediction model, the external environment data is obtained, and the external environment data is input into the preset compensation confirmation model to obtain the real-time score of the external environment data.

[0072] External environmental data can be data on external factors that affect process parameters such as thermal process, temperature control, air flow distribution and pressure. Specifically, it can include meteorological data: such as temperature, humidity, wind speed, air pressure, etc. Plant environment data: such as indoor temperature, humidity, air quality, wind speed, etc. Equipment status information: such as the load of the area where the equipment is located, power supply status, equipment operation status, etc. External airflow data: such as the flow rate and air pressure of fans and air handling equipment.

[0073] The preset compensation confirmation model can be a trained machine learning model that is designed to determine whether thermal process parameter (pressure, gas flow, temperature, etc.) compensation is required based on external environmental data. The core purpose of this model is to determine whether the control parameters in the furnace should be adjusted based on changes in the external environment to ensure wafer quality and process stability.

[0074] The real-time score can be a value output by the model after inputting external environment data into the compensation confirmation model, which indicates the degree of influence of the external environment on the thermal process technology. The score is usually a continuous value or a discrete level, which is used to determine whether the current external environment is extreme enough and whether the process parameters in the furnace need to be adjusted.

[0075] External environmental data can be obtained in real time through sensors, weather stations or other monitoring equipment. These data include temperature, humidity, air pressure, wind speed, equipment status, plant environment, etc. The acquired external environmental data is input into the compensation confirmation model. The model will calculate a real-time score based on these input data. Specifically, the scoring branch of the model can analyze the external environmental data and output a score value, which can be a single value.

[0076] S203, if the real-time score does not exceed the preset score threshold, there is no need to perform pressure compensation, gas flow compensation and temperature compensation, and the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube.

[0077] The preset scoring threshold can be a pre-set scoring threshold based on historical data, process requirements, and the degree of influence of external environmental changes. When the real-time score exceeds this threshold, it means that the change in the external environment is sufficient to affect the stability of the thermal process and compensation is required; otherwise, it can be considered that the change in the external environment has little impact on the process and the current process parameters are maintained.

[0078] The real-time score can be compared with a preset score threshold. A decision is made based on the result of the comparison: If the real-time score does not exceed the score threshold: it means that the external environment has little impact on the heat treatment process, and no pressure, gas flow or temperature compensation is required. When pressure, gas flow or temperature compensation is not required, the step of "sending the pressure adjustment value to the pressure control system of each preset control zone in each furnace tube, sending the gas flow adjustment value to the gas flow control valve of each preset control zone in each furnace tube, and sending the temperature adjustment value to the temperature control system of each preset control zone in each furnace tube" can be executed.

[0079] Based on the above technical solution, optionally, after obtaining the real-time score of the external environment data, the method further includes:

[0080] If the real-time score exceeds a preset score threshold, a pressure compensation value, a gas flow compensation value, and a temperature compensation value are output through a preset compensation confirmation model;

[0081] Correspondingly, the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube, including:

[0082] The pressure adjustment value and the pressure compensation value are sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value and the gas flow compensation value are sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value and the temperature compensation value are sent to the temperature control system of each preset control zone in each furnace tube.

[0083] In this solution, the pressure compensation value can be an adjustment value of the furnace pressure based on external environmental changes (such as pressure changes, temperature changes, etc.). Its purpose is to compensate for the furnace pressure deviation caused by external environmental fluctuations and ensure that the furnace pressure remains within the optimal working range.

[0084] The gas flow compensation value can be an adjustment value of the gas flow based on external environmental changes (such as humidity, wind speed, etc.). Its purpose is to ensure that the gas flow in the furnace can be maintained at the expected level when the external environment changes, ensuring the stability of the thermal process.

[0085] The temperature compensation value can be an adjustment value of the furnace temperature according to the external temperature change (such as seasonal temperature difference, atmospheric temperature change, etc.). Its purpose is to compensate for the impact of external temperature changes on the thermal process in the furnace and ensure that the furnace temperature is maintained within a predetermined range.

[0086] The current external environment data (temperature, humidity, air pressure, etc.) can be obtained in real time through the external environment sensor. These real-time external environment data are input into the preset compensation confirmation model, and the scoring branch of the model will calculate a real-time scoring value, which reflects the impact of the external environment on the current thermal process. The preset scoring threshold is used to determine whether the current change in the external environment requires compensation adjustment. If the real-time score exceeds the threshold, it means that the external environment fluctuation has a greater impact on the thermal process and compensation adjustment is required; otherwise, no compensation is required. If the real-time score exceeds the preset scoring threshold, the compensation confirmation model will output the corresponding compensation value, pressure compensation value: the compensation calculation branch of the model will output the pressure compensation value calculated according to the external environment conditions. Gas flow compensation value: the compensation calculation branch of the model will output the gas flow compensation value calculated according to the external environment conditions. Temperature compensation value: the compensation calculation branch of the model will output the temperature compensation value calculated according to the external environment conditions. Then the calculated pressure compensation value and the real-time adjustment value are sent to the pressure control system in each furnace tube through wireless communication technology to adjust the gas pressure in the furnace. The gas flow compensation value and the real-time adjustment value are sent to the gas flow control valve in each furnace tube through wireless communication technology to adjust the gas flow. The temperature compensation value and the real-time adjustment value are sent to the temperature control system in each furnace tube through wireless communication technology to adjust the temperature in the furnace.

[0087] In this solution, the implementation of a real-time compensation strategy based on external environmental data not only helps to improve the stability of the thermal process and reduce uncertainty in production, but also optimizes the use of resources, reduces costs, automates the production process, and improves production efficiency and product quality.

[0088] Based on the above technical solution, the optional training process of the preset compensation confirmation model includes:

[0089] Acquire a historical fault scoring record, determine first historical external environment data and a historical score according to the historical fault scoring record, and mark a scoring tag of the first historical external environment data according to the historical score;

[0090] Obtain a historical compensation record, determine the second historical external environment data, the historical pressure compensation value, the historical gas flow compensation value, and the historical temperature compensation value according to the historical compensation record, and mark the pressure compensation value label of the second historical external environment data according to the historical pressure compensation value, mark the gas flow compensation value label of the second historical external environment data according to the historical gas flow compensation value, and mark the temperature compensation value label of the second historical external environment data according to the historical temperature compensation value;

[0091] Constructing a compensation confirmation model, and training a scoring branch of the compensation confirmation model according to the first historical external environment data and the historical scores;

[0092] The compensation calculation branch of the compensation confirmation model is trained according to the second historical external environment data, the pressure compensation value label, the gas flow compensation value label and the temperature compensation value label.

[0093] In this solution, historical scoring records may refer to past scoring data, which records the scoring results under specific external conditions. Scoring may be based on certain criteria (such as equipment operating status, external environmental conditions) to assess whether compensation processing is required.

[0094] The first historical external environment data may refer to external environment data (such as temperature, air pressure, humidity, etc.) collected in past historical records, and these data are used to evaluate the impact of the external environment on the device or system.

[0095] The historical score can be a score obtained by evaluating the historical external environment data. It is usually a quantitative indicator that indicates how well the external environment data performs under a certain standard.

[0096] Scoring labels can be a way to associate historical scores with external environment data for training models. Each piece of historical external environment data is assigned a scoring label to indicate whether compensation is needed in that environment and the intensity of the compensation.

[0097] The historical compensation records may include records of compensation performed in the past, such as the execution time of the compensation process, and the compensation value (pressure, gas flow, temperature, etc.). These data are used to determine the compensation effect.

[0098] The second historical external environment data may be similar to the first historical external environment data, which refers to another set of historically collected external environment data. The difference is that the second set of data will be associated with a specific compensation value to guide the compensation calculation.

[0099] The historical pressure compensation value, the historical gas flow compensation value, and the historical temperature compensation value may be the compensation values ​​of the various control variables (pressure, gas flow, and temperature) when the compensation operation is performed in the historical records. The compensation value indicates the adjustment value performed in order to maintain a certain target state.

[0100] The pressure compensation value tag may be a tag in a historical pressure compensation record, used to mark the pressure compensation value required under specific external environment data.

[0101] The gas flow compensation value label can be used to mark the gas flow compensation value required under specific external environment.

[0102] The temperature compensation value tag may be a compensation value tag related to temperature, indicating the temperature compensation amount required under a specific external environment.

[0103] The scoring branch can be a branch in the compensation confirmation model, which is trained using historical external environmental data and scores, with the purpose of determining whether compensation is needed, as well as the type and intensity of compensation based on different environmental scores.

[0104] The compensation calculation branch can be another branch of the compensation confirmation model, which is trained using the second historical external environment data and compensation value labels (pressure compensation, gas flow compensation, temperature compensation). The goal of this branch is to predict and calculate the required compensation value based on the external environment data.

[0105] All historical scoring records can be extracted from the data storage or log, and the external environmental data (such as temperature, air pressure, etc.) corresponding to each record can be determined. At the same time, the historical scoring labels are used to evaluate whether these environmental data need to be compensated. The external environmental data (such as pressure, gas flow, temperature) and the corresponding compensation values ​​(pressure compensation, gas flow compensation, temperature compensation) at that time are obtained from the historical compensation records. The scoring labels are marked through the first historical external environmental data and historical scores; the pressure, gas flow, and temperature compensation value labels are marked through the second historical external environmental data and historical compensation records. Using the first historical external environmental data and historical scores to train the scoring branch, the model will learn how to determine whether compensation is needed based on the external environmental data. Using the second historical external environmental data, pressure compensation value labels, gas flow compensation value labels, and temperature compensation value labels, the compensation calculation branch is trained, and the model will learn how to calculate the actual required compensation value based on the external environmental data and compensation value labels.

[0106] Third embodiment

[0107] The third embodiment of the present application relates to a method for reducing the increment of wafer warpage. The third implementation is an improvement on the first embodiment, and the specific improvement is that it includes the following steps:

[0108] S301, obtaining the real-time pressure, real-time gas flow and real-time temperature of each preset control area in each furnace tube, and obtaining the property data of the wafer that needs to be subjected to the thermal process, and inputting the real-time pressure, real-time gas flow, real-time temperature and property data into the preset warping prediction model to obtain the warping prediction value of the wafer that needs to be subjected to the thermal process.

[0109] S302, if the warpage prediction value exceeds the preset warpage threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube are output through the preset warpage prediction model, and the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube.

[0110] S303, obtaining operation status data and equipment performance data of the furnace body, inputting the operation status data and the equipment performance data into a preset fault detection model, and obtaining a fault score of the furnace body.

[0111] The furnace body may be a device for thermally treating wafers (eg, silicon wafers in a semiconductor manufacturing process).

[0112] Operation status data can refer to real-time data of the furnace during operation, specifically, current / voltage: current and voltage information of the electric heating system. Atmosphere control: for example, gas composition and oxygen concentration in the furnace. Vibration: mechanical vibration data inside the furnace, which may indicate mechanical failure.

[0113] Equipment performance data can be the long-term performance data of the furnace body. Specifically, it can include service life: the time and cycle of equipment use. Historical fault records: including equipment maintenance records, fault types, and repair time. Maintenance records: equipment maintenance and maintenance history. Parts wear: such as the wear of heating elements, gas flow control valves, etc.

[0114] The preset fault detection model can be a model based on machine learning or statistical methods, which is used to analyze the operating status and performance data of the equipment and predict whether the equipment has a risk of failure.

[0115] The fault score can be calculated by the fault detection model to measure the current fault risk or health of the furnace. The score is usually a numerical value, the higher the probability of the equipment failure, the lower the score, the better the equipment is running.

[0116] Real-time operating status data and equipment performance data can be obtained through sensors and monitoring systems installed on the furnace body. This data is then input into the fault detection model for processing and analysis. The scoring branch of the fault detection model outputs a fault score based on the input data, indicating the current health status or fault risk of the furnace body. The calculated fault score is then compared with the preset fault score threshold. If the fault score is lower than or equal to the threshold, it means that the furnace body is operating normally and production can continue. If the fault score exceeds the threshold, it means that the furnace body may have a fault or a risk of failure, and may require maintenance or shutdown inspection. If the fault score does not exceed the threshold, the system sends a continue operation instruction to the furnace control system to allow the equipment to continue operating.

[0117] S304: If the fault score does not exceed a preset fault score threshold, a continue operation instruction is sent to the furnace body.

[0118] The preset fault scoring threshold can be a critical value used to determine whether the device needs to take action. Usually, a scoring threshold is set based on the historical data of the device, production requirements, and safety standards.

[0119] The continue operation instruction may be an instruction issued by the system when the equipment is in good operating condition and the fault score does not exceed a preset threshold, instructing the furnace to continue production operation.

[0120] If the fault score does not exceed the preset fault score threshold, a continue operation instruction is sent to the furnace body through wireless communication technology.

[0121] Based on the above technical solution, optionally, after obtaining the fault score of the furnace body, the method further includes:

[0122] If the fault score exceeds a preset fault score threshold, the fault type is output through a preset fault detection model and the fault type is sent to a control center.

[0123] In this solution, the fault type may refer to a specific problem or abnormality that occurs in the furnace body during the thermal process. The fault type helps to identify the source of the furnace failure and provide guidance for subsequent processing. Specifically, it may include temperature abnormality: the temperature inside the furnace body deviates from the predetermined range, which may be caused by a heating element failure, a temperature control system failure, or a sensor failure. Pressure abnormality: The pressure in the furnace body exceeds the set value, which may be caused by a pressure regulation system failure or poor sealing. Gas flow abnormality: The gas flow rate is not within the predetermined range, which may be due to a gas flow sensor failure, insufficient gas supply, or a flow control system failure. Equipment mechanical failure: The mechanical parts of the furnace body are damaged, such as a damaged furnace door, a heating element falling off, a bracket failure, etc. Electrical failure: Including power system failure, electrical failure of the sensor or control board. Chemical reaction problems: For example, abnormal oxygen or nitrogen concentrations in the atmosphere may affect the effect of wafer processing.

[0124] The control center can refer to a centralized management system responsible for monitoring, managing and directing factories, equipment, production processes, etc. The control center is generally equipped with technical facilities for data monitoring, fault diagnosis, alarm and decision support.

[0125] The fault scoring branch of the fault detection model can be used to analyze the operating status data and equipment performance data of the furnace body to obtain a fault score. If this fault score exceeds the preset fault score threshold, it indicates that the furnace body may have a fault. In the case where the fault score exceeds the preset threshold, the fault type classification branch of the fault detection model outputs the fault type based on the historical data and the results of the training model. The fault type analyzes the relationship between the fault score and the historical score, combined with the current operating status of the equipment, to infer the possible fault category. Once the fault type is determined, the fault type data can be sent to the control center through wireless communication technology. The control center can arrange for the repair or replacement of the faulty component to prevent further equipment damage or production interruption.

[0126] In this solution, through real-time monitoring and fault scoring, potential faults of the furnace can be identified in time, so that preventive measures can be taken before the fault occurs to avoid the occurrence of serious faults.

[0127] On the basis of the above technical solution, optionally, the training process of the preset fault detection model includes:

[0128] Acquire historical fault score records, determine historical operation status data, historical equipment performance data, and a first historical fault score of the furnace body according to the historical fault score records, create a third data set according to the historical operation status data and the historical equipment performance data, and annotate a fault score label of the third data set according to the first historical fault score;

[0129] Determine a first historical fault score that exceeds a preset fault score threshold as a second historical fault score, obtain a historical fault type of the second historical fault score, create a fourth data set with historical operating status data and historical equipment performance data corresponding to the second historical fault score, and annotate a fault type label of the fourth data set according to the historical fault type;

[0130] Constructing a fault detection model, and training a fault scoring branch of the fault detection model according to the third data set and the fault scoring label;

[0131] The fault type classification branch of the fault detection model is trained according to the fourth data set and the fault type label.

[0132] In this solution, the historical fault scoring record can be a scoring record of the fault status of the equipment or furnace during its past operation. These scores reflect the severity of the equipment failure and are usually generated by a monitoring system or a fault diagnosis system. The scores may be based on the equipment's operating data, sensor data, or preset fault judgment rules.

[0133] Historical operating status data may include various monitoring data of the equipment during operation, such as temperature, pressure, load, current, vibration, etc., which reflects the operating environment and status of the equipment.

[0134] The historical equipment performance data may be long-term performance data of the equipment, such as efficiency, response time, load change, etc., indicating whether the equipment is within a normal operating range or whether performance degradation has occurred.

[0135] The first historical fault score may be a score value obtained according to historical operating status data and equipment performance data.

[0136] The third data set can be a data set constructed based on historical operating status data and historical equipment performance data. Each piece of data contains the equipment's operating status information and equipment performance indicators. This data set will be used to train the fault scoring branch, with the goal of allowing the model to understand the performance of the equipment under different operating conditions and learn how to predict the equipment's fault score.

[0137] The fault score label may be a label corresponding to each piece of data in the third data set, indicating a fault score of the device.

[0138] The second historical fault score can be the records that exceed the preset fault score threshold selected from the historical fault score records. Scores exceeding the threshold indicate that a significant fault has occurred in the device or the severity of the fault exceeds the normal range. These data are used for more detailed fault analysis to help the system determine the specific fault type of the device.

[0139] The historical fault type may refer to the fault type corresponding to the second historical fault score. It is usually determined through fault diagnosis, expert system or system log, and may be a mechanical fault, electrical fault, sensor fault, etc. Understanding the fault type is the basis for subsequent fault classification.

[0140] The fourth data set can be a data set constructed based on the second historical fault score and the corresponding historical fault type and equipment operation data. Each piece of data not only contains the equipment operation data, but also comes with the fault type label of the fault score record. This data set is used to train the fault type classification branch, allowing the model to learn how to classify different types of faults based on the historical data of the equipment.

[0141] The fault type label may be a label corresponding to each piece of data in the fourth data set, used to identify the type of fault. It may include different fault types such as mechanical fault, electrical fault, overtemperature, abnormal pressure, etc. These labels are used to train the fault type classification branch so that the model can predict different fault types.

[0142] The fault scoring branch can be a submodule of the fault detection model, which aims to predict the fault score of the equipment based on the historical operating status and performance data of the equipment. It learns through training how to give a numerical value based on the input data to indicate the severity of the equipment failure.

[0143] The fault scoring branch can be a submodule of the fault detection model, which aims to predict the fault score of the equipment based on the historical operating status and performance data of the equipment. It learns through training how to give a numerical value based on the input data to indicate the severity of the equipment failure.

[0144] Historical fault scoring records can be obtained from the fault monitoring system or log. These records contain scoring information for each equipment fault, reflecting the operating status and fault severity of the equipment. From the historical fault scoring records, extract the historical operating status data and equipment performance data, and annotate each record with a fault scoring label (such as a fault scoring value). Create a third data set using the historical operating status data and equipment performance data, and attach a fault scoring label to each data. These data are used to train the fault scoring branch of the model. From the historical fault scoring records, filter out the fault scores that exceed the preset threshold and identify more serious fault events. For each historical fault score that exceeds the threshold, determine and record the fault type, such as mechanical fault, electrical fault, etc. This information is used for subsequent fault type classification. Create a fourth data set using the second historical fault score and the corresponding historical fault type. This data set contains the operating data and performance data of the equipment and the corresponding fault type label. Then create a fourth data set using the second historical fault score and the corresponding historical fault type. This data set contains the operating data and performance data of the equipment and the corresponding fault type label. After the model training is completed, the fault detection model can predict its fault score based on the real-time operating data of the equipment and further identify the fault type.

[0145] The step division of the above methods is only for clear description. When implemented, they can be combined into one step or some steps can be split and decomposed into multiple steps. As long as they include the same logical relationship, they are all within the protection scope of this application; adding insignificant modifications to the algorithm or process or introducing insignificant designs without changing the core design of the algorithm and process are all within the protection scope of this patent.

[0146] Fourth embodiment

[0147] A fourth embodiment of the present application relates to a system for reducing the increment of wafer warpage variation, and the system specifically includes the following:

[0148] The warpage prediction module 401 is used to obtain the real-time pressure, real-time gas flow rate and real-time temperature of each preset control area in each furnace tube, and obtain the property data of the wafer to be subjected to the thermal process, and input the real-time pressure, real-time gas flow rate, real-time temperature and property data into a preset warpage prediction model to obtain the warpage prediction value of the wafer to be subjected to the thermal process;

[0149] The adjustment value confirmation module 402 is used to output the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube through a preset warping prediction model if the warping prediction value exceeds the preset warping threshold value, and send the pressure adjustment value to the pressure control system of each preset control zone in each furnace tube, send the gas flow adjustment value to the gas flow control valve of each preset control zone in each furnace tube, and send the temperature adjustment value to the temperature control system of each preset control zone in each furnace tube.

[0150] It is not difficult to find that this embodiment is a system embodiment corresponding to the first embodiment, and this embodiment can be implemented in conjunction with the first embodiment. The relevant technical details mentioned in the first embodiment are still valid in this embodiment, and in order to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in this embodiment can also be applied in the first embodiment.

[0151] It is worth mentioning that all modules involved in this embodiment are logic modules. In practical applications, a logic unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, in order to highlight the innovative part of this application, this embodiment does not introduce units that are not closely related to solving the technical problems proposed by this application, but this does not mean that there are no other units in this embodiment.

[0152] The flow chart or block diagram in the accompanying drawings shows the possible architecture, function and operation of the equipment, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated system for hardware that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0153] The scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes that fall within the meaning and scope of the equivalent elements of the claims are included in the present application. Any figure mark in the claims should not be regarded as limiting the claims involved. In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or systems stated in the system claim may also be implemented by one unit or system through software or hardware. The words "first", "second", etc. are only used to distinguish the description, and do not indicate any specific order, nor can they be understood as indicating or implying relative importance.

[0154] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily mention changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims, and the above embodiments should be regarded as exemplary and non-restrictive.

Claims

1. A method for reducing the incremental change in wafer warpage, characterized in that: The method comprises: Acquire the real-time pressure, real-time gas flow rate and real-time temperature of each preset control area in each furnace tube, and acquire the property data of the wafer to be subjected to the thermal process, and input the real-time pressure, real-time gas flow rate, real-time temperature and property data into the preset warpage prediction model to obtain the warpage prediction value of the wafer to be subjected to the thermal process; If the warping prediction value exceeds the preset warping threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube are output through the preset warping prediction model, and the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube.

2. The method for reducing the increment of wafer warpage according to claim 1, characterized in that: After sending the gas flow adjustment value to the gas flow control valve of each preset control zone in each furnace tube, the method further includes: Real-time monitoring to determine whether a preset collection time point has been reached. If the preset collection time point has been reached, the real-time pressure, real-time gas flow rate, and real-time temperature of each preset control area in each furnace tube are re-acquired, and the real-time pressure, real-time gas flow rate, real-time temperature, and property data are re-input into a preset warpage prediction model to update the warpage prediction value. If the updated warpage prediction value exceeds the preset warpage threshold, the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube are re-output through the preset warpage prediction model, and the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube until the thermal process of the wafer is completed.

3. The method for reducing the increment of wafer warpage variation according to claim 1, characterized in that: After outputting the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube through the preset warpage prediction model, the method further includes: Acquire external environment data, input the external environment data into a preset compensation confirmation model, and obtain a real-time score of the external environment data; If the real-time score does not exceed the preset score threshold, there is no need to perform pressure compensation, gas flow compensation and temperature compensation.

4. The method for reducing the increment of wafer warpage according to claim 1, characterized in that: After obtaining the real-time score of the external environment data, the method further includes: If the real-time score exceeds a preset score threshold, a pressure compensation value, a gas flow compensation value, and a temperature compensation value are output through a preset compensation confirmation model; Correspondingly, the pressure adjustment value is sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value is sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value is sent to the temperature control system of each preset control zone in each furnace tube, including: The pressure adjustment value and the pressure compensation value are sent to the pressure control system of each preset control zone in each furnace tube, the gas flow adjustment value and the gas flow compensation value are sent to the gas flow control valve of each preset control zone in each furnace tube, and the temperature adjustment value and the temperature compensation value are sent to the temperature control system of each preset control zone in each furnace tube.

5. The method for reducing the wafer warpage variation increment according to claim 1, characterized in that: After sending the temperature adjustment value to the temperature control system of each preset control zone in each furnace tube, the method further includes: Acquire operation status data and equipment performance data of the furnace body, input the operation status data and the equipment performance data into a preset fault detection model, and obtain a fault score of the furnace body; If the fault score does not exceed the preset fault score threshold, a continue operation instruction is sent to the furnace body.

6. The method for reducing the wafer warpage variation increment according to claim 5, characterized in that: After obtaining the fault score of the furnace body, the method further includes: If the fault score exceeds a preset fault score threshold, the fault type is output through a preset fault detection model and the fault type is sent to a control center.

7. The method for reducing wafer warpage variation increment according to claim 1, characterized in that: The training process of the preset warpage prediction model includes: Obtaining historical thermal process processing records, and determining historical pressure, historical gas flow, historical temperature, historical property data of wafers to be subjected to thermal process processing, and a first historical warpage value of each preset control zone in each furnace tube according to the historical thermal process processing records; Determine a first historical warpage value exceeding a preset warpage threshold as a second historical warpage value, and obtain a historical pressure adjustment value, a historical gas flow adjustment value, and a historical temperature adjustment value of each preset control zone in each furnace tube corresponding to the second historical warpage value; Creating a first data set according to the historical pressure, historical gas flow, historical temperature of each preset control zone in each furnace tube and the historical property data of the wafer to be subjected to the thermal process, and labeling the first data set with a warpage value label according to the first historical warpage value; Creating a second data set according to the historical pressure, historical gas flow, historical temperature of each preset control zone in each furnace tube corresponding to the second historical warpage value and the historical property data of the wafer to be subjected to the thermal process, marking the pressure label of the second data set according to the historical pressure, marking the gas flow label of the second data set according to the historical gas flow, and marking the temperature label of the second data set according to the historical temperature; Constructing a warpage prediction model, and training a warpage prediction branch of the warpage prediction model according to the first data set and the warpage value label; The adjustment value calculation branch of the warpage prediction model is trained according to the second data set, the pressure label, the gas flow label, and the temperature label.

8. The method for reducing wafer warpage variation increment according to claim 3, characterized in that: The training process of the preset compensation confirmation model includes: Acquire a historical fault scoring record, determine first historical external environment data and a historical score according to the historical fault scoring record, and mark a scoring tag of the first historical external environment data according to the historical score; Obtain a historical compensation record, determine the second historical external environment data, the historical pressure compensation value, the historical gas flow compensation value, and the historical temperature compensation value according to the historical compensation record, and mark the pressure compensation value label of the second historical external environment data according to the historical pressure compensation value, mark the gas flow compensation value label of the second historical external environment data according to the historical gas flow compensation value, and mark the temperature compensation value label of the second historical external environment data according to the historical temperature compensation value; Constructing a compensation confirmation model, and training a scoring branch of the compensation confirmation model according to the first historical external environment data and the historical scores; The compensation calculation branch of the compensation confirmation model is trained according to the second historical external environment data, the pressure compensation value label, the gas flow compensation value label and the temperature compensation value label.

9. The method for reducing the wafer warpage variation increment according to claim 5, characterized in that: The training process of the preset fault detection model includes: Acquire historical fault score records, determine historical operation status data, historical equipment performance data, and a first historical fault score of the furnace body according to the historical fault score records, create a third data set according to the historical operation status data and the historical equipment performance data, and annotate a fault score label of the third data set according to the first historical fault score; Determine a first historical fault score that exceeds a preset fault score threshold as a second historical fault score, obtain a historical fault type of the second historical fault score, create a fourth data set with historical operating status data and historical equipment performance data corresponding to the second historical fault score, and annotate a fault type label of the fourth data set according to the historical fault type; Constructing a fault detection model, and training a fault scoring branch of the fault detection model according to the third data set and the fault scoring label; The fault type classification branch of the fault detection model is trained according to the fourth data set and the fault type label.

10. A system for reducing the incremental change in wafer warpage, characterized in that: The system comprises: A warpage prediction module is used to obtain the real-time pressure, real-time gas flow rate and real-time temperature of each preset control area in each furnace tube, and to obtain the property data of the wafer to be subjected to the thermal process, and input the real-time pressure, real-time gas flow rate, real-time temperature and property data into a preset warpage prediction model to obtain the warpage prediction value of the wafer to be subjected to the thermal process; The adjustment value confirmation module is used to output the pressure adjustment value, gas flow adjustment value and temperature adjustment value of each preset control zone in each furnace tube through a preset warping prediction model if the warping prediction value exceeds a preset warping threshold value, and send the pressure adjustment value to the pressure control system of each preset control zone in each furnace tube, send the gas flow adjustment value to the gas flow control valve of each preset control zone in each furnace tube, and send the temperature adjustment value to the temperature control system of each preset control zone in each furnace tube.

Images