A semiconductor process sampling detection method

Abstract

The application relates to the technical field of semiconductors, and particularly discloses a semiconductor process sampling detection method, which comprises the following steps: configuring a to-be-detected station according to a product main flow and establishing an association relationship between the detection station and an upstream process station of the detection station; matching an equipment type according to the process station and configuring a sampling detection rule; recording information of the process station after a batch is processed in the process station; when the batch arrives at the detection station for the first time, sampling detection is carried out, and information of the detection station is recorded; if the batch arrives at the detection station for the first time, a difference between maximum cumulative information of the batch in the process station and the recorded information of the detection station is calculated, and when the difference is greater than or equal to a set threshold value, the batch is sampled and detected; the detection station performs wafer selection, if the selected wafers are not enough, candidate wafers are screened according to wafer information of the detection station, and the number of the wafers is supplemented. Through dynamic configuration of the association relationship between the detection station and the process station, the sampling frequency and the number can be accurately and flexibly controlled, and the detection efficiency is significantly improved.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and in particular to a semiconductor process sampling detection method. Background Art

[0002] A wafer is a silicon chip used in the manufacture of silicon semiconductor integrated circuits. Because of its circular shape, it is called a wafer. Various circuit component structures can be fabricated on the silicon wafer, resulting in IC products with specific electrical functions. During the mass production process of wafers in a wafer fab, wafers are measured and received at measurement stations. When production volumes are high, comprehensive testing of every wafer is unrealistic. Wafers must be inspected during the manufacturing process. If every product were to be inspected, numerous SCAN scanners would be required, which would be quite costly and would also affect production efficiency. Therefore, sampling inspection is typically used.

[0003] The current method is to conduct random inspections based on the last digit of the product batch number or the number of products on the machine. The existing random inspection method is very simple and has problems such as low flexibility and accuracy, and low detection efficiency. Summary of the Invention

[0004] To overcome the aforementioned technical problems existing in the prior art, an embodiment of the present invention provides a semiconductor process sampling inspection method. By dynamically configuring the relationship between inspection sites and process sites, and setting inspection conditions based on batch, wafer quantity, time, and combination rules, this method enables precise and flexible control of inspection frequency and quantity. By recording key data such as wafer processing time and cumulative quantity, combined with candidate wafer screening and patching mechanisms, sampling coverage is optimized. For multi-sub-machine scenarios, independent inspection rule configuration and priority coordination are supported, significantly improving inspection efficiency and process monitoring accuracy.

[0005] In order to achieve the above-mentioned purpose, an embodiment of the present invention provides a semiconductor process sampling inspection method, comprising the following steps: S1: configuring the site to be inspected according to the main process of the product, and establishing an association relationship between the inspection site and its upstream process site; S2: matching the corresponding equipment type according to the processing capacity of the process site, and configuring sampling rules, the sampling rules including at least one sampling type of batch sampling, wafer sampling, time sampling, and combined sampling; S3: after the batch is processed at the process site, the process site records the associated inspection site information and the first wafer information, the first wafer information including the cumulative batch number, the cumulative wafer number and the time when the wafer leaves the equipment; S4: when the batch arrives at the inspection site for the first time, the site is inspected, and the inspection site records the associated process site, equipment name and second wafer information, the The second wafer information includes the cumulative number of inspection batches, the number of wafers entering the station, and the time information of the last wafer in the batch leaving the equipment; S5: If the batch does not arrive at the inspection site for the first time, calculate the difference between the maximum cumulative wafer information of the batch at the process site and the latest recorded information of the inspection site. When the difference is greater than or equal to the set threshold of the inspection type, the batch enters the station for inspection; S6: The inspection site performs a wafer selection operation to determine whether the target number of wafers is fully selected within the current inspection cycle. If not, the candidate wafers are screened according to the wafer information in the batch and the wafer information in the historical records of the inspection site, and the remaining number is supplemented. If the number of candidate wafers exceeds the target number, they are randomly selected; S7: When the process site contains multiple sub-machines, each sub-machine independently configures the inspection rules, and gives priority to wafers that meet the most inspection rules in the inspection site.

[0006] Preferably, the batch sampling inspection is specifically as follows: setting each completed N batches as a sampling inspection cycle; the wafer sampling inspection is specifically as follows: setting each processed M wafers as a sampling inspection cycle; the time sampling inspection is specifically as follows: setting each interval time T as a sampling inspection cycle; the combined sampling inspection is specifically as follows: setting the sampling inspection conditions according to at least two parameter combinations of batch, wafer, and time; the sampling inspection rules further include: when configured for combined sampling inspection, the station will be inspected if any set threshold is met, or the station will be inspected after all set thresholds are met.

[0007] Preferably, the wafer selection operation in step S6 specifically includes: if the wafer selection is not full, recording the candidate wafer information and performing patching in subsequent batches, and selecting wafers in the patch batch that are larger than the maximum wafer information of the previous batch as patches.

[0008] Preferably, step S2 also includes: configuring a virtual measurement site after the process site, and using the virtual measurement site to perform measurement operations on the batch; obtaining measurement results, and when the measurement results do not exceed specifications / exceed control, executing steps S3-S7; when the measurement results are beyond specifications / out of control / no measurement results, the batch is inspected at the in-station, and a random inspection wafer rule is configured and random inspection is performed, and the random inspection wafer rule is as follows: if all wafers in the batch have no measurement results, random inspection is performed according to a preset number of wafers and the batch is inspected at the in-station; if all wafers in the batch are measured, and the measurement results of some wafers are beyond specifications or out of control, the batch is inspected at the in-station; if there are unmeasured wafers in the batch, random inspection is performed from the unmeasured wafers according to a preset number of wafers and the batch is inspected at the in-station.

[0009] Preferably, the method also includes: when the number of process sites in step S1 is greater than 1, each process site is associated with the detection relationship, each process site is configured according to step S2, and each process site records information according to step S3; when the batch arrives at the detection site, if one of the process sites meets the set threshold, the batch enters the site for random inspection, if each process site does not meet the set threshold, the batch does not enter the site for random inspection; after the batch enters the detection site, the detection site executes step S4 according to the association relationship; if all process sites meet the batch entry random inspection conditions, the wafer selection operation is performed according to step S6 according to the association relationship, and the wafers covering the most processing equipment are randomly inspected; if one of the process sites meets the batch entry random inspection conditions, and wafers are processed at other process sites, the detection site updates the accumulated information according to the process site.

[0010] Preferably, the method further comprises: recording sampling events: recording batch information, process site information, sampling steps to reach the inspection site, and reasons for entering or skipping the site.

[0011] Preferably, the method further includes: configuring sampling inspection parameters, setting the priority of the sampling inspection parameters, performing logical judgment according to the priority, obtaining a judgment result, and triggering entry or station skipping or other sampling inspection rules according to the judgment result.

[0012] Preferably, the sampling inspection parameters include a reference queue time threshold, an incoming batch number, a skipped batch number, a small batch wafer quantity threshold, and a number of wafers to be sampled.

[0013] Preferably, the priority of the sampling parameters is set, specifically including sorting the priority of the sampling parameters in order from large to small as follows: reference queue time threshold, entry batch number, skip station batch number, small batch wafer quantity threshold, and sampling wafer quantity.

[0014] The technical solution provided by the present invention has at least the following technical effects:

[0015] By dynamically configuring the relationship between inspection sites and process sites, and setting sampling conditions based on batch, wafer quantity, time and combination rules, the frequency and quantity of sampling can be accurately and flexibly controlled; by recording key data such as wafer processing time and cumulative quantity, combined with candidate wafer screening and patching mechanisms, the sampling coverage rate can be optimized; for multi-sub-machine scenarios, independent sampling rule configuration and priority coordination are supported, significantly improving inspection efficiency and process monitoring accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present invention, but do not constitute a limitation of the embodiments of the present invention. In the accompanying drawings:

[0017] Figure 1 This is a flow chart of a semiconductor process sampling detection method provided by an embodiment of the present invention;

[0018] Figure 2 It is a structural diagram of a semiconductor process sampling detection system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following describes the specific implementation of the embodiment of the present invention in detail with reference to the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the embodiment of the present invention and is not used to limit the embodiment of the present invention.

[0020] The terms "system" and "network" in the embodiments of the present invention can be used interchangeably. "Multiple" refers to two or more. In view of this, "multiple" can also be understood as "at least two" in the embodiments of the present invention. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / ", unless otherwise specified, generally indicates that the previous and next associated objects are in an "or" relationship. In addition, it should be understood that in the description of the embodiments of the present invention, words such as "first" and "second" are only used to distinguish the purpose of description, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order.

[0021] See Figure 1 , an embodiment of the present invention provides a semiconductor process sampling detection method, comprising the following steps:

[0022] S1: Configure the sites to be tested according to the main product process, and establish the association between the testing sites and their upstream process sites;

[0023] S2: Matching the corresponding equipment type according to the processing capability of the process site, configuring sampling rules, wherein the sampling rules include at least one sampling type of batch sampling, wafer sampling, time sampling, and combination sampling;

[0024] S3: After the batch is processed at the process site, the process site records the associated inspection site information and first wafer information, where the first wafer information includes the cumulative batch number, the cumulative number of wafers, and the time when the wafers leave the equipment;

[0025] S4: When the batch arrives at the inspection site for the first time, it is inspected and the inspection site records the associated process site, equipment name, and second wafer information. The second wafer information includes the cumulative number of inspection batches, the number of wafers in the incoming batch, and the time when the last wafer in the batch left the equipment.

[0026] S5: If the batch does not arrive at the inspection site for the first time, calculate the difference between the maximum cumulative wafer information of the batch at the process site and the latest recorded information at the inspection site. When the difference is greater than or equal to the set threshold of the sampling inspection type, the batch enters the site for sampling inspection.

[0027] S6: The inspection station performs a wafer selection operation to determine whether the target number of wafers has been selected in the current sampling cycle. If not, candidate wafers are screened based on the wafer information in the batch and the wafer information in the inspection station's historical records, and the remaining number is supplemented. If the number of candidate wafers exceeds the target number, random selection is performed;

[0028] S7: When the process site includes multiple sub-tools, each sub-tool is independently configured with the sampling inspection rules, and wafers that meet the most sampling inspection rules are preferentially selected in the inspection site.

[0029] In the embodiment of the present invention, in a mass production environment, especially in the field of semiconductor and precision manufacturing, batches are processed from different process stations, such as cleaning, etching, deposition, lithography, grinding, etc., and the process of arriving at the inspection station is a key link. Therefore, certain parameters can be configured to perform target wafer sampling (abbreviated as Target Sampling), where the parameters for configuring sampling inspection include: process processing capability, combination rules, inspection type, number of selected wafers, total number of wafers, total number of batches and time, among which process processing capability is used to match the process processing capability to the corresponding equipment; combination rules are used to select the corresponding equipment type, and the equipment types include: equipment (EQP), chamber (Chamber), interface component or connector (Header), container for storing and processing chemical liquids (Tank, such as the trough used in wet etching or cleaning process), chuck (Chuck, a device for clamping and positioning wafers); the number of selected wafers is the number of wafers that need to be inspected at the inspection site; the total number of wafers is used to set the number of wafers in a cycle; the total number of batches is used to set the number of batches in a cycle; and the time is used to set the duration of a cycle. There are many types of mass production wafer production processes and the process processes are also different. In some scenarios, the wafer reaches the inspection site after passing through one process site, and in some scenarios, it reaches the inspection site after passing through multiple process sites.

[0030] Specifically, according to the main process configuration under the product, the site that needs to be inspected is configured, which is the inspection site. Then, the process site upstream of the inspection site is configured, and an association relationship is established between the inspection site and the process site. According to the processing capacity of the process site, multiple equipment for random inspection is matched. Then, according to the process of the process site, the required equipment type is selected for random inspection, and random inspection rules are configured for the selected equipment type. The random inspection rules include at least one of the following inspection types, including batch inspection, wafer inspection, time inspection, and combined inspection.

[0031] In one embodiment, the sampling inspection type can be formulated according to the following rules: batch sampling inspection: set every 3 completed batches as a sampling inspection cycle; wafer sampling inspection: set every 20 wafers processed as a sampling inspection cycle; time sampling inspection: set every 2 hours as a sampling inspection cycle; combined sampling inspection: set the sampling inspection conditions based on a combination of at least two parameters among batch, wafer, and time; the specific setting parameters can be adjusted according to the factory's production plan to ensure more accurate and flexible sampling inspection.

[0032] Furthermore, when a batch of wafers is processed at the manufacturing site, the process site will record the following information: the associated inspection site, the cumulative number of processing batches, the cumulative number of wafers processed at the process site, and the time when the wafers leave the equipment, where the cumulative number of processing batches is obtained by accumulating the batches processed at the site, and the cumulative number of wafers is obtained by accumulating the number of wafers from all batches processed at the process site; the batch of wafers arrives at the associated inspection site, and the inspection site determines whether the batch arrives at this site for the first time. If it arrives for the first time, the batch of wafers enters the inspection site for random inspection, and the inspection site records the associated process site information, equipment The equipment name information is collected and random inspections are carried out according to the random inspection rules configured for this equipment name. In addition, the inspection site will also record the cumulative number of random inspection batches, the number of wafers in the random inspection batches entering the station, and the time when the wafers leave the equipment. The cumulative number of random inspection batches is obtained by accumulating the number of batches inspected at the station, and the number of wafers in the random inspection batches entering the station is obtained by refreshing the number of wafers in the batch inspected at the time. If it is not the first arrival, calculate the difference between the maximum cumulative value of the wafer information recorded at the process site where the wafers of this batch are located and the latest wafer information recorded at the associated inspection site, and determine whether the difference is greater than or equal to the set threshold of the random inspection type. If it is greater than or equal to, the batch is inspected at the station. Inspection, if it is less than, the batch will skip the station; if the sampling inspection rule configured by the equipment type is the combined sampling inspection type, then when the batch entering the station for sampling inspection meets the preset value of one of the sampling inspection types in the combined sampling inspection or meets the preset values ​​of all sampling inspection types, the batch can enter the station for sampling inspection; for example, a factory configures the inspection site according to its own product main process, and finds the process site upstream of the inspection site, and configures the association relationship between the process site and the inspection site. The number of wafers in a batch is 20. After the batch is processed at the process site, it arrives at the associated inspection site. The sampling inspection rule configured by the matching equipment type is the wafer sampling inspection type. The specific rule is to set the wafer sampling inspection to the same level every 20 wafers processed. If a circle is a sampling inspection cycle, the set threshold of the sampling inspection type is 20; the cumulative batch wafer quantity of the process site before processing the batch is 30, and after processing the batch, the cumulative batch wafer quantity of the process site is refreshed to 50. The batch arrives at the inspection site associated with the process site, and this batch is not the first time to arrive at this inspection site. The inspection site records the number of wafers of the incoming sampling inspection batch as 10 before the batch arrives. After the batch arrives at the inspection site, the number of wafers of the incoming sampling inspection batch is refreshed to 20. The calculated difference is 30 (50-20=30). The calculated difference 30 is greater than or equal to the set threshold 20, so the batch enters the station for sampling inspection.

[0033] Furthermore, after the batch enters the station for random inspection, the inspection site must perform the wafer selection operation. First, it determines whether the wafer selection for the inspection cycle is full based on the configured wafer selection quantity parameter (that is, the number of wafers that the inspection site needs to inspect). If it is full, the inspection site will record the wafer selection information as 0, and compare each wafer in the batch with the maximum wafer information in the last batch that entered the station recorded by the inspection site. If there is a wafer in this batch that is greater than the maximum information, the wafer will be included in the candidate wafer. If the number of candidate wafers is greater than the number of wafers that need to be inspected, random wafers will be selected from the candidate wafers to select the final wafer. As the sampling wafer; if the number of wafers is not full, the selection information recorded by the inspection site will show the number of selected wafers, and each wafer in this batch will be compared with the maximum information of the wafers in the last batch entering the station recorded by the inspection site. The wafers in the batch with information greater than the maximum information will be used as candidate wafers, and the last wafer will be selected from the candidate wafers as the sampling wafer. If the number of candidate wafers does not reach the number of wafers that need to be selected, all wafers in the candidate wafers will be selected, and the patching operation will be performed in the next batch entering the station for inspection; finally, the inspection site will be used to inspect the sampling wafers to obtain the inspection results.

[0034] For the patching operation, in one embodiment, when the next incoming inspection batch enters the inspection site, the information of each wafer in this batch is compared with the maximum wafer information in the previous incoming batch at the inspection site. If there is a wafer in this batch that is larger than the maximum wafer information, the wafer is used as a patching candidate, and the number of patches required is selected from the patching candidates to complete the patching operation. Finally, the inspection operation is performed to obtain the sampling results.

[0035] Furthermore, when a process site uses multiple sub-machines to process a batch, each sub-machine is configured with a separate sampling rule. When the batch arrives at the inspection site associated with the process site, the candidate wafers of each sub-machine are first selected, and then the wafers that meet the most sampling types among these candidate wafers are preferentially selected as sampling wafers. Finally, the sampling wafers are inspected using the inspection site to obtain the sampling results.

[0036] Through the semiconductor process sampling inspection method provided by this embodiment, different process sites can select corresponding inspection sites, and different sampling rules can be configured according to each device. The sampling rules also include multiple sampling types, and each sampling type is configured with corresponding rules to control the number of wafers sampled and the frequency of sampling, thereby improving the accuracy and flexibility of sampling inspection.

[0037] In one embodiment, after establishing the associated process sites and inspection sites in steps S1 to S2 and configuring them, first configure a virtual measurement site after the process site. The site type of the virtual measurement site is a measurement site, and the processing capacity of the site is a virtual measurement machine. The virtual measurement site can calculate the measurement results of the wafers through big data; the batch is measured using the virtual measurement site. After the measurement, if the measurement results issued by the virtual measurement site are out of specification or out of control or there is no measurement result for all wafers, the batch is inspected at the inspection site that has a management relationship with the process site, and the configured entry-in random inspection wafer rule is: if all wafers in the batch have no measurement values ​​at the virtual measurement site, all wafers processed at the process site are inspected at the inspection site, and the number of random inspection wafers is equal to the configured number of selected wafers. If the batch If the number of wafers in the batch is not enough, no wafers will be added and the configuration cycle will not be checked; if all wafers in the batch have been measured at the virtual measurement site, but the measurement results of only some wafers are out of specification or out of control, only those wafers that are out of specification or out of control need to be sampled and sent to the inspection site for inspection. The number of sampled wafers is equal to the number of configured selected wafers. If the number of wafers is not enough, no wafers will be added and the configuration cycle will not be checked; if only some wafers in the batch have been measured at the virtual measurement site, then for other wafers that have not been measured, find these wafers that have not been measured and sample them from the wafers that have not been measured. The number of sampled wafers is equal to the number of configured selected wafers. If the number of wafers is not enough, no wafers will be added and the configuration cycle will not be checked; if all wafers in the batch have been measured at the virtual measurement site, and the measurement results are not out of specification or out of control or there is no small number of wafers, perform operations according to steps S3-S7.

[0038] The sampling inspection method provided in this embodiment adopts virtual measurement combined with random inspection. By using the big data of the virtual machine at the virtual measurement site to calculate the wafer value results, it is only necessary to conduct random inspections on unqualified wafers at the inspection site. The virtual site and virtual machine replace the physical object, thereby improving the quality and accuracy of random inspections.

[0039] In one embodiment, after establishing the associated process sites and inspection sites in steps S1 to S2 and configuring them, a pre-measurement site is configured before the process site, and the site type of the pre-measurement site is a measurement site; when a batch arrives at the pre-measurement site, the pre-measurement site will enter the site for measurement according to the measurement and sampling rules configured at this site, and record the measured wafers; when the batch arrives at the inspection site, it is first determined according to steps S3-S7 whether the batch meets the requirements. If it meets the requirements, it is determined whether the wafers of the batch have been measured at the pre-measurement site. If they have been measured, the batch enters the site, and the wafers for sampling are the wafers measured at the pre-measurement site; if they have not been measured at the pre-measurement site, the batch skips the inspection site and does not enter the site for inspection. The inspection method provided by this embodiment adopts pre-measurement combined with sampling inspection, and determines the wafers for sampling inspection at the inspection site based on the wafers measured at the pre-measurement site, so as to more effectively detect the quality of the wafers and improve the efficiency of sampling inspection.

[0040] In one embodiment, when the process site is a multi-process, that is, when the number of process sites is greater than 1, the association relationship between each process site and the detection site is established respectively through step S1, and each process site is configured through step S2, and each process site performs recording information according to step S3; when the batch arrives at the detection site, if one of the process sites meets the set threshold of the sampling type configured for the site, the batch enters the site for sampling inspection; if all the process sites do not meet the set threshold of the configured sampling type, the batch skips the detection site and does not enter the site for sampling inspection; through the above judgment, when the batch enters the detection site, the detection site executes step S4 according to each association record, that is, for each association, the cumulative number of wafers entering the site, the cumulative number of batches, and the number of wafers are recorded at the detection site. The equipment time information is output, and the inspection site performs the wafer selection operation according to step S6 based on each association relationship, and gives priority to sampling out the wafers with the most covered equipment; if the batch arrives at the inspection site after being processed by multiple process sites, and all processes meet the wafer information for batch entry inspection, then the respective wafers are pre-selected according to the association relationship between each process site and the inspection site, and finally the wafer with the most processing equipment is preferentially selected from the pre-selected wafers as the inspection wafer; if the batch arrives at the inspection site after being processed by multiple process sites, and the wafers sampled by one of the process sites meet the number of wafers that need to be sampled and there are wafers that have been processed on the equipment of other process sites, then these wafers are used as inspection wafers, and the inspection site updates the cumulative information of the same wafers of other processes at the inspection site. Through the sampling inspection method provided in this embodiment, combined with multiple processes, wafers processed on multiple devices are preferentially inspected at the inspection site. The advantage is that more time can be saved, and the wafers do not need to be sent to different processes and then to the inspection site for inspection. Especially in the mass production process, this technical solution can save more time costs and improve production efficiency.

[0041] In actual application, during wafer mass production, the same process sites and measurement sites exist in the main process of the same product or between the main processes of different products. Frequent spot checks are carried out under the same site process, which will greatly reduce production efficiency. This is also a pain point in the business. The current solution is for the same main process under the same product or different main processes under different products. The processing capabilities between process sites are the same, the equipment used for process sites is consistent, and the processing capabilities between inspection sites are the same. When a batch of processing is completed at one of the process sites and arrives at the inspection site, the associated process site information and inspection site information are refreshed accordingly. This is called "mutual brushing". By configuring the number of mutual brushings while ensuring product quality, a process in the package is refreshed by other processes multiple times in a row, which will result in this process not being actually measured in multiple spot check cycles.

[0042] In order to solve the above technical problems, in one embodiment, two or more association relationships are bound as package inspection, and are configured according to steps S1 to S2, and the setting conditions for package inspection are: the processing capabilities of the process sites are the same, the processing capabilities of the inspection sites are the same, and the equipment processed by the process sites are the same. The package inspection is also configured with a parameter named inspection cycle count, which is used to set the maximum number of inspections to be mutually brushed. After the maximum number is reached, the batch needs to be processed at the process site and arrive at the inspection site for forced inspection of wafers; when the process sites and the inspection sites are different under the same process for the same product, the specific process is as follows: (1) When the batch is processed at one of the process sites, the site information record is made according to step S3, and the wafers are brushed synchronously (1) When the batch arrives at one of the inspection stations, determine whether the batch is to be inspected for wafers according to the relationship between the process station and the inspection station and according to steps S4 to S7; (2) Then refresh the other inspection stations. Specifically, if the batch can be fully inspected at the current inspection station, then refresh the information of other inspection stations; if the batch is fully inspected at the current inspection station, then refresh the information of other inspection stations; if the inspection station is refreshed all the time and the number of refreshes reaches the maximum number of inspection cycle counts, then the batch must be inspected at the inspection station; when the process stations are different and the inspection stations are different under different products and processes, the above-mentioned processes (1) to (3) are also followed.

[0043] The sampling inspection method provided in this embodiment solves the problem that when the processing capabilities of the process sites and the inspection sites in the production line are the same, there is no need to conduct inspections on each process site. At the same time, when the processing capabilities of the process sites and the inspection sites are the same under different products and different processes, there is no need to send each batch of wafers to the site for random inspection. Through the packaged random inspection method, a certain process may be repeatedly brushed by other processes, which will result in the process not being actually measured within multiple random inspection cycles. This embodiment sets a bottom-picking number of times. When the number of brushings exceeds the set maximum value, the random inspection rule is that the batch is forced to enter the random inspection site. The advantage of this is to ensure the consistency of product quality and performance. The sampling exchange method is used for quality control and testing. This approach is an effective means to improve production efficiency and product reliability, and avoid errors in the measurement data of the wafers inspected.

[0044] In actual application, there is an important random inspection in the production process. This process involves random inspection of samples in the furnace to ensure that the materials and components used in the production process meet strict quality standards. The furnace tube random inspection is mainly used to monitor and control the temperature, atmosphere and other key parameters in the wafer production process. Through random inspection measurement and analysis, it can be ensured that the conditions in the furnace are suitable for the growth and processing of wafers, thereby ensuring the quality and consistency of the final product. In the furnace tube random inspection, how to ensure that the batches inspected by technicians will cover the wafers in every area of ​​the furnace tube equipment.

[0045] In order to solve the above technical problems, in one embodiment, the site to be inspected is configured according to the main process of the product, and the process site before the inspection site is configured according to the configured inspection site, and the processing equipment is a furnace tube equipment, wherein the configured parameters also include batch group cycle (referred to as Group Batch Size, used to calculate how many batch groups constitute a cycle), selection of batches to enter the station for each set area, and selection of batches for polling in the set area. Three parameter information; when a batch is processed on the furnace tube equipment of the process site, the batch entering the station for random inspection will be marked, and when the batch arrives at the inspection site, it will be determined whether it needs to enter the station for inspection based on the label on the batch; the specific steps for marking the batch are as follows: because there are two ways to conduct furnace tube random inspection, when the furnace tube random inspection method is to select a batch in each area of ​​the furnace tube equipment within a cycle, first, it is taken as a cycle according to the configured batch group cycle parameters, and then when the batch is processed at the process site, it is marked for the batch selected in the area. Batches are marked. If there are areas in the furnace tube equipment that have not been selected for a batch within a cycle, a new batch will be selected in each area for marking in the next cycle. When the marked batch arrives at the inspection station and enters the station, the wafers are randomly inspected and tested according to the configured selection rules. When the furnace tube sampling method is to select a batch from one area in the furnace tube equipment within a cycle, the configured batch group cycle parameters are first used as a cycle. Then, when the batch is processed at the process station, the batch in the selected area is marked. Then, the area in the furnace tube equipment is polled according to the cycle to select batches. After the marked batch arrives at the inspection station and enters the station, the wafers are randomly inspected and tested according to the configured selection rules. The method for determining the furnace tube area is as follows: first, the area to which the batch belongs in the furnace tube equipment is determined by the number of wafers. If the number of wafers is the same, the area to which the batch belongs in the furnace tube equipment is determined based on the order of the slot positions of the wafers in the area.

[0046] Through the sampling inspection method provided in this embodiment, for random inspections within the furnace tube, the batches inspected will cover the wafers in every area of ​​the furnace tube equipment. The random inspection coverage is wide, thereby ensuring the quality and consistency of the final product. During the furnace tube inspection, technicians will extract a batch from different areas inside the furnace for measurement and analysis. The results will help engineers adjust parameters in the production process, such as temperature, gas flow and pressure, to optimize wafer quality. The inspection data can also be used for fault analysis and continuous improvement of production processes.

[0047] In actual production, especially in wafer fabs focused on high precision and efficiency, determining the sampling ratio calculation cycle is crucial. The calculation cycle (or sampling frequency) refers to the time interval between key parameter checks during the production process. A correct sampling cycle not only ensures product quality but also effectively controls production costs and improves production efficiency. Because products often involve precision manufacturing and demand extremely high quality, the sampling ratio is typically adjusted based on the process stability of the product flow. This allows for timely detection of any deviations in the production process, enabling timely adjustments to avoid further material waste and production delays, ensuring that each production step meets the highest precision standards.

[0048] However, excessively high sampling frequencies can also lead to problems, such as increased data processing burdens and higher production costs. Therefore, when determining the sampling frequency, a balance must be struck between ensuring quality and controlling costs. This typically requires experimentation and historical data analysis to determine the optimal sampling frequency. The sampling frequency in wafer production is a complex but crucial decision, directly impacting product quality and production efficiency. By scientifically calculating and adjusting the sampling frequency, it is possible to optimize production processes, reduce costs, and improve competitiveness while ensuring product quality.

[0049] The system calculates the sampling ratio and then the real-time cycle only for Target Sampling and Fixed Sampling + Target Sampling configurations. The first configuration parameters include the number of equipment that has passed the test within 14 days (Active Chamber Count), the average number of equipment that has passed the test within 14 days (Average Run Chamber Count), and the lot size (Lot Size). The Active Chamber Count is calculated based on the configuration rule group dimension by adding the number of non-duplicate equipment that has passed the test within 14 days at the process site with the same first four digits of the product name. The Average Run Chamber Count is calculated using the function: Equipment that has passed the test within 14 days / Total Equipment Number (Equipment that has passed the test within 14 days refers to non-duplicate EQPs that have passed the test within 14 days at the process site with the same first four digits of the product name). The lot size is uniformly 23.

[0050] Specifically, the calculation formula is as follows: 1. When the sampling inspection type is time sampling inspection, use formula (1) to calculate, specifically:

[0051] Calculate the sampling ratio = Number of devices that pass within 14 days * Number of wafers selected * 24 hours * Batch size * 30 / time * Data collection specification configuration Number of wafers selected * Capacity

[0052] 2. When the sampling inspection type is batch sampling, use formula (2) to calculate, specifically:

[0053] Calculate the sampling ratio = average number of devices that pass within 14 days * number of selected wafers / total number of batches * number of wafers selected based on data collection specifications

[0054] 3. When the sampling inspection type is wafer sampling inspection, use formula (3) to calculate, specifically:

[0055] Calculate the sampling ratio = batch size * number of selected wafers / total number of wafers * number of wafers selected by data collection specification configuration

[0056] In one embodiment, after the random inspection is completed, a random inspection event recording operation must be performed. The random inspection event record mainly records batch information, process site information, random inspection steps to reach the inspection site, and reasons for entering or jumping the station. To this end, a way is provided to track and record the specific reasons for entering or jumping the inspection site. It can also track and record the wafers selected by the station. This is a very important function. This information is crucial for analyzing the causes, impacts and consequences of the incident. Through the recorded random inspection events, technicians can trace the random inspection process from a specific batch and a specific process site to the inspection site through the event records, and can effectively perform data analysis.

[0057] Based on the same inventive concept, see Figure 2 An embodiment of the present invention provides a semiconductor process sampling detection system, the sampling detection system comprising:

[0058] Configuration module: used to set the relationship between inspection sites and process sites and the sampling inspection rules;

[0059] Data recording module: used to store the cumulative number of batches, the cumulative number of wafers and the time when the wafers leave the equipment;

[0060] Random inspection judgment module: Based on random inspection rules and comparison of historical data, it determines whether the batch should be inspected;

[0061] Wafer selection execution module: performs wafer selection and patching operations according to candidate wafer screening rules;

[0062] Multi-device coordination module: When there are multiple sub-machines, it integrates the sampling results of each sub-machine and optimizes the film selection priority.

[0063] In the actual production process, each product has different measurement standards for different processes. For random inspection of specific batches in the same batch, a specified sampling (Fixed Sampling) can be performed. In one embodiment, the following sampling parameters are first configured: reference queue time threshold, entry batch number, skip batch number, small batch wafer quantity threshold and sampling wafer quantity; wherein the reference queue time threshold (abbreviated as RQT) is represented by the remaining maximum Q-Time, and Q-Time represents the waiting time of the batch wafers between different process steps; entry batch number is represented by the batch code for setting the direct entry to the station upon arrival; skip batch number is represented by the batch code for setting the direct skip to the station upon arrival; small batch wafer quantity threshold: whether the number of wafers in the set batch meets the entry requirement, if not, it will skip directly; sampling wafer quantity: represents the number of wafers that need to be sampled at the inspection station; and these sampling parameters are configured with a priority order, which is sorted from high to low in order of priority: reference queue time threshold>entry batch number>skip batch number>small batch wafer quantity threshold>spotted wafer quantity (abbreviated as WaferSelect Count); perform logical judgment in order of priority. The priority logical judgment is as follows:

[0064] 1. When a batch arrives at the testing station, the Q-Time of the batch is first compared with the configured RQT. If the Q-Time of the batch is less than the configured RQT, the batch will skip the station. If the Q-Time of the batch is greater than the configured RQT, the calibration will continue.

[0065] 2. Then verify the incoming batch number. If the batch number meets the incoming batch number, the incoming wafers in this batch will be randomly inspected, and the number of wafers sampled will be the configured Wafer Select Count. If the batch number does not meet the incoming batch number, continue to verify the batch skip station number.

[0066] 3. If the batch number meets the batch skip station number, the batch will skip the station; if the batch number does not meet the batch skip station number, the small batch wafer quantity threshold will be checked;

[0067] 4. If the number of wafers in a batch is less than the configured small batch wafer quantity threshold, the batch will be skipped; if the number of wafers in a batch is greater than the small batch wafer quantity threshold, other configured sampling inspection methods will continue to be verified. If no other sampling inspection method is configured, the batch will be skipped.

[0068] Based on the above priority logic judgment, it can be determined whether the batches arriving at the testing site should enter the station for testing. Specific batches within the same batch can be randomly tested and designated to enter the station for measurement. If the batches within the same batch pass the random inspection, the remaining batches will be skipped to reduce the measurement and save time costs.

[0069] In practical applications, if the production process of wafers using specific batch numbers can be effectively controlled and optimized during mass production, it can ensure that product quality meets the highest standards. In one embodiment, the parameters are adjusted according to the specific needs of the batch, and two parameters, namely, the default tail number and the designated code, are added on the basis of the parameter settings for the above-mentioned designated sampling inspection. The default tail number parameter is the tail number determined by adjusting the ratio, and the designated code is used to select the tail number of the designated batch number. When the batch arrives at the inspection site, the process steps 1 to 4 of the above-mentioned priority logic judgment process are first used to determine whether the batch meets the conditions. If the conditions are not met, the batch tail number is checked. If the batch tail number meets the tail number of the designated batch number, the batch enters the station for random inspection of wafers. If the tail number of the designated batch number is not met, the batch is skipped. The method provided in this embodiment can effectively control the production process of wafers using specific batch numbers. It is particularly suitable for high-end products or key change points that may occur in the production process. It helps to discover and solve problems in a timely manner and ensure the stable operation of the entire production line.

[0070] The above describes in detail the optional implementation methods of the embodiments of the present invention in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above implementation methods. Within the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the scope of protection of the embodiments of the present invention.

[0071] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any appropriate manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe various possible combinations.

[0072] Those skilled in the art will understand that all or part of the steps in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a program. The program is stored in a storage medium and includes a number of instructions for causing a single-chip microcomputer, chip, or processor to execute all or part of the steps in the methods described in the various embodiments of the present application. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0073] In addition, various implementations of the embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the embodiments of the present invention, they should also be regarded as the contents disclosed in the embodiments of the present invention.

Claims

1. A semiconductor process sampling detection method, characterized in that: The following steps are involved: S1: Configure the sites to be tested according to the main product process, and establish the association between the testing sites and their upstream process sites; S2: Matching the corresponding equipment type according to the processing capability of the process site, configuring sampling rules, wherein the sampling rules include at least one sampling type of batch sampling, wafer sampling, time sampling, and combination sampling; S3: After the batch is processed at the process site, the process site records the associated inspection site information and first wafer information, where the first wafer information includes the cumulative batch number, the cumulative number of wafers, and the time when the wafers leave the equipment; S4: When the batch arrives at the inspection site for the first time, it is inspected and the inspection site records the associated process site, equipment name, and second wafer information. The second wafer information includes the cumulative number of inspection batches, the number of wafers in the incoming batch, and the time when the last wafer in the batch left the equipment. S5: If the batch does not arrive at the inspection site for the first time, calculate the difference between the first wafer information of the batch at the process site and the latest recorded value of the second wafer information of the batch at the inspection site. When the difference is greater than or equal to a set threshold of the sampling inspection type, the batch enters the site for sampling inspection. S6: The inspection station performs a wafer selection operation to determine whether the target number of wafers has been selected in the current sampling cycle. If not, candidate wafers are screened based on the wafer information in the batch and the wafer information in the inspection station's historical records, and the remaining number is supplemented. If the number of candidate wafers exceeds the target number, random selection is performed; S7: When the process site includes multiple sub-tools, each sub-tool is independently configured with the sampling inspection rules, and wafers that meet the most sampling inspection rules are preferentially selected in the inspection site.

2. A semiconductor process sampling detection method according to claim 1, characterized in that: In step S2, the batch sampling inspection is specifically as follows: setting each completed N batches as a sampling inspection cycle; The wafer sampling inspection is specifically as follows: setting each processing of M wafers as a sampling inspection cycle; The time sampling inspection is specifically as follows: setting each interval time T as a sampling inspection cycle; The combined random inspection specifically includes: setting random inspection conditions based on at least two parameters of batch, wafer, and time; The sampling inspection rules further include: when configured as combined sampling inspection, the station will be inspected if any set threshold is met, or the station will be inspected if all set thresholds are met.

3. A semiconductor process sampling detection method according to claim 2, characterized in that: The wafer selection operation in step S6 specifically includes: if the wafer selection is not full, recording the candidate wafer information and performing wafer patching in a subsequent batch, and selecting a wafer in the patch batch that is larger than the maximum wafer information in the previous batch as a patch.

4. The semiconductor process sampling detection method according to claim 3, characterized in that: The step S2 further includes: configuring a virtual measurement site after the process site, and using the virtual measurement site to perform measurement operations on the batch; Obtaining measurement results, and when the measurement results are within specifications / out of control, executing steps S3-S7; When the measurement result is out of specification / out of control / no measurement result, the batch enters the station for inspection, configures the random inspection wafer rules and performs random inspection. The random inspection wafer rules are as follows: If there are no measurement results for all wafers in the batch, a preset number of wafers will be randomly inspected and sent to the station for inspection; If all wafers in the batch are measured and the measurement results of some wafers are out of specification or out of control, the wafers are randomly sampled for incoming inspection; If there are unmeasured wafers in the batch, a preset number of wafers are randomly selected from the unmeasured wafers and sent to the inspection station for inspection.

5. The semiconductor process sampling detection method according to claim 3, characterized in that: The method further includes: when the number of process sites in step S1 is greater than 1, establishing an association relationship between each process site and a detection relationship, configuring each process site according to step S2, and recording information for each process site according to step S3; When a batch arrives at a testing station, if one of the process stations meets the set threshold, the batch will be sent to the station for random inspection; if each process station does not meet the set threshold, the batch will not be sent to the station for random inspection; After the batch enters the testing site, the testing site executes step S4 according to the association relationship; If all process sites meet the batch incoming sampling inspection conditions, the wafer selection operation is performed according to the association relationship according to step S6, and the wafers covering the most processing equipment are sampled; If one of the process sites meets the batch entry sampling inspection conditions and wafers are processed at other process sites, the inspection site updates the accumulated information based on the process site.

6. A semiconductor process sampling detection method according to any one of claims 3 to 5, characterized in that: The method further includes: recording sampling inspection events: recording batch information, process site information, sampling inspection steps to the inspection site, and reasons for entering or skipping the site.

7. The semiconductor process sampling detection method according to claim 1, wherein: The method further includes: configuring sampling inspection parameters, setting the priority of the sampling inspection parameters, performing logical judgment according to the priority, obtaining a judgment result, and triggering entry or station skipping or other sampling inspection rules according to the judgment result.

8. The semiconductor process sampling detection method according to claim 7, characterized in that: The sampling inspection parameters include a reference queue time threshold, an incoming batch number, a skipped batch number, a small batch wafer quantity threshold, and a number of wafers to be sampled.

9. The semiconductor process sampling detection method according to claim 8, characterized in that: The priority of the sampling parameters is set, specifically, the priority of the sampling parameters is sorted in order from large to small as follows: reference queue time threshold, entry batch number, skip station batch number, small batch wafer quantity threshold, and sampling wafer quantity.

Images