CN115705543A - Method and apparatus for evaluating throughput of semiconductor device - Google Patents
Method and apparatus for evaluating throughput of semiconductor device Download PDFInfo
- Publication number
- CN115705543A CN115705543A CN202110924495.7A CN202110924495A CN115705543A CN 115705543 A CN115705543 A CN 115705543A CN 202110924495 A CN202110924495 A CN 202110924495A CN 115705543 A CN115705543 A CN 115705543A
- Authority
- CN
- China
- Prior art keywords
- data
- queue
- group
- acquiring
- effective
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- General Factory Administration (AREA)
Abstract
The embodiment of the application relates to the field of semiconductor manufacturing, and provides a method and a device for evaluating the production capacity of semiconductor equipment, wherein the method comprises the following steps: acquiring the time interval between the end of the current batch production and the end of the previous batch secondary production of the current batch of all the process machines; acquiring effective data in the data as effective data; acquiring at least two initial data sets; acquiring a first data set based on the initial data set; sorting the data in the first data group to form a first queue data group, and acquiring a standard deviation of the data in the first queue data group; obtaining a second queue data set based on the first queue data set and the standard deviation; acquiring minimum effective data and maximum effective data based on data in all the second queue data groups; and acquiring the production capacity of the semiconductor equipment based on the minimum effective data and the maximum effective data. The embodiment of the application is beneficial to improving the accuracy of the evaluated production capacity of the semiconductor equipment.
Description
Technical Field
The embodiment of the application relates to the field of semiconductor manufacturing, in particular to a method and a device for evaluating the production capacity of semiconductor equipment.
Background
One of the tasks of capacity planning in a semiconductor manufacturing plant is to plan the number of semiconductor devices to be produced according to the capacity expansion demand, and estimate the investment amount according to the planned number of semiconductor devices. The number of semiconductor devices is mainly determined by the production capacity of the semiconductor devices, and the production capacity of the semiconductor devices indirectly determines the investment amount required for capacity expansion.
Currently, in the semiconductor manufacturing field, when a semiconductor device is fully loaded and stably operated, it is customary to measure the throughput of the semiconductor device (WPH) by the number of wafers produced per unit time for a certain semiconductor device and a certain process step. The mathematical definition of WPH is: WPH = Run Size/Takt Time, where Run Size indicates the maximum number of wafers allowed to be produced for each lot of a certain semiconductor device, run Size is a fixed value when the semiconductor device is in a fully loaded state, and Takt Time indicates the Time interval between the end of the current lot of production and the end of the previous lot of production of the process tool in the semiconductor device.
However, as the semiconductor device is operated, the semiconductor device is not always in a fully loaded and stable operation state at every moment, and there is a difference between Takt times collected at different moments, so that there is a difference between WPHs at different moments, and these factors interfere with the evaluation of the production capacity of the semiconductor device. Therefore, a method for evaluating the throughput of semiconductor devices is needed to improve the accuracy of the evaluated throughput of semiconductor devices.
Disclosure of Invention
The embodiment of the application provides a method and a device for evaluating the production capacity of semiconductor equipment, which are at least beneficial to improving the accuracy of the evaluated production capacity of the semiconductor equipment.
According to some embodiments of the present application, in one aspect, there is provided a method for evaluating the throughput of a semiconductor device, the semiconductor device including different types of process tools, the method comprising: producing and processing N batches of wafers by using all the process machines, and acquiring data of all the process machines, wherein the data is a time interval between the end of the current batch production of the process machines and the end of the previous batch secondary production of the current batch; acquiring effective data in the data as effective data; acquiring at least two initial data sets, wherein the data in each initial data set is the effective data of each batch when the same type of process machine platform produces the same product and carries out the same process step; acquiring a first data set based on the initial data set, wherein data in the first data set is the effective data corresponding to the working state of all chambers in the process machine table in the initial data set; sorting the data in the first data group to form a first queue data group, and acquiring a standard deviation of the data in the first queue data group; acquiring a second queue data group based on the first queue data group and the standard deviation, wherein the second queue data group is formed by sorting data left after removing the next data in the previous and next data when the difference value of the adjacent previous and next data in the first queue data group deviates from the standard deviation; acquiring minimum effective data and maximum effective data based on data in all the second queue data groups; and acquiring the production capacity of the semiconductor equipment based on the minimum effective data and the maximum effective data.
According to some embodiments of the present application, another aspect of the embodiments of the present application further provides an apparatus for evaluating the production capacity of a semiconductor device, comprising: the data collection module is used for acquiring data of all process machines, wherein the data is the time interval between the end of the production of the current batch of the process machines and the end of the production of the previous batch of the current batch; a data processing module for processing the data, the data processing module configured to: acquiring effective data in the data as effective data; acquiring at least two initial data sets, wherein the data in each initial data set is the effective data of each batch when the same type of process machine table produces the same product and carries out the same process step; acquiring a first data set based on the initial data set, wherein data in the first data set is the effective data corresponding to the working state of all chambers in the process machine table in the initial data set; sorting the data in the first data group to form a first queue data group, and acquiring a standard deviation of the data in the first queue data group; acquiring a second queue data group based on the first queue data group and the standard deviation, wherein the second queue data group is formed by sorting data left after removing the next data in the previous and next data when the difference value of the adjacent previous and next data in the first queue data group deviates from the standard deviation; and the acquisition module is used for acquiring minimum effective data and maximum effective data based on the data in all the second queue data groups and acquiring the production capacity of the semiconductor equipment based on the minimum effective data and the maximum effective data.
Compared with the related art, the technical scheme provided by the embodiment of the application has the following advantages:
in the technical scheme, after data (namely Takt Time, time interval between the end of the current batch production of the process machine and the end of the previous batch secondary production of the current batch) of all process machines are obtained, effective data are screened from the data to improve the accuracy of the obtained effective data, and at least two initial data sets are obtained from the effective data; then screening effective data in the initial data set to obtain a first data set so as to improve the accuracy of the effective data in the first data set; then, ordering the data in the first data group to obtain a first queue data group, so as to facilitate other subsequent processing on the effective data and obtain the standard deviation of the data in the first queue data group; and screening the effective data in the first queue data group to obtain a second queue data group based on the standard deviation so as to improve the accuracy of the effective data in the second queue data group. Therefore, the accuracy of the effective data in the finally formed second queue data group is improved by screening the acquired data of all the process machines for at least three times. Therefore, when the minimum effective data and the maximum effective data are acquired based on the data in all the second queue data sets subsequently and the production capacity of the semiconductor equipment is further acquired, the accuracy of the estimated production capacity of the semiconductor equipment is improved by improving the accuracy of the acquired minimum effective data and the acquired maximum effective data, so that the accuracy of the investment amount required by capacity expansion estimation is improved.
Drawings
One or more embodiments are illustrated by corresponding figures in the drawings, which are not to scale unless otherwise specified.
FIG. 1 is a flowchart illustrating a method for evaluating the throughput of a semiconductor device according to an embodiment of the present application;
FIG. 2 is another detailed flow chart of a method for evaluating the throughput of a semiconductor device according to an embodiment of the present application;
FIG. 3 is a flowchart illustrating a method for evaluating the throughput of a semiconductor device according to an embodiment of the present application
Fig. 4 is a functional block diagram of an apparatus for evaluating throughput of a semiconductor device according to another embodiment of the present application.
Detailed Description
As is known in the art, the accuracy of the current estimated throughput of semiconductor devices is subject to improvement.
Analysis shows that the Run Size is a fixed value when the semiconductor equipment is in a full-load state, and the equipment production capacity is determined by Takt Time when the semiconductor equipment is in a full-load and stable-operation state.
Because of the extreme precision of semiconductor device fabrication, takt Time generally has two main manifestations: (1) The Takt Time is controlled by a clock in the semiconductor equipment, and for a determined processing step, the Takt Time has smaller fluctuation range along with the increase of processing batch; (2) The Takt Time is controlled by an advanced process control system in the semiconductor equipment, and is continuously corrected by feeding back key parameters of a process in the process of acquiring the Takt Time, wherein the Takt Time is in a slow increasing trend along with the increase of processing batches.
Therefore, under the condition that the semiconductor equipment is fully loaded and stably runs, the Takt Time has the maximum value and the minimum value, namely the Takt Time has the boundary value, and further the maximum value and the minimum value also exist in the normal production capacity of the semiconductor equipment. In order to obtain a valid boundary for Takt Time, it is necessary to correctly process Takt Time. At present, the processing of Takt Time is mainly described by a normal distribution algorithm, and the concept of boundary values is not involved. However, since Takt Time in semiconductor manufacturing has its unique distribution characteristics, normal distribution can only roughly describe the average level of Takt Time, and the accuracy of determining the average level of Takt Time is not high, and the accuracy of evaluating the production capacity of semiconductor devices using the average level of Takt Time further decreases, and the lower the accuracy of evaluating the production capacity of semiconductor devices, the greater the deviation of investment amount required for capacity planning is generated. The investment required for capacity planning in the semiconductor industry is typically on the order of billions of dollars, with even one percent deviation resulting in hundreds of millions of dollars investment gap.
In the evaluation method, the accuracy of effective data in a finally formed second queue data group is improved by screening acquired data (Takt Time) of all process machines for at least three times. Therefore, when the minimum effective data and the maximum effective data, that is, the boundary value of the Takt Time, are obtained subsequently based on the data in all the second queue data sets, the accuracy of the finally evaluated semiconductor device production capacity is improved by improving the accuracy of the obtained minimum effective data and the obtained maximum effective data. In addition, because the semiconductor manufacturing process flow is complex and the production line condition is dynamically variable, the determination of the boundary value of the Takt Time can play an important role, for example, the maximum value and the minimum value of the production capacity of the semiconductor equipment are determined through the boundary value of the Takt Time, a solid decision basis is provided for the capacity planning of a semiconductor manufacturing factory, and the accuracy of the determined maximum value and the determined minimum value of the investment amount is favorably improved through improving the accuracy of the obtained boundary value of the Takt Time, so that the decision efficiency is improved on one hand, and the accuracy of the boundary value of the investment amount required by capacity expansion evaluation is improved on the other hand.
To make the objects, technical solutions and advantages of the embodiments of the present application clearer, the embodiments of the present application will be described in detail below with reference to the accompanying drawings. However, it will be appreciated by those of ordinary skill in the art that in the examples of the present application, numerous technical details are set forth in order to provide a better understanding of the present application. However, the technical solution claimed in the present application can be implemented without these technical details and various changes and modifications based on the following embodiments.
An embodiment of the present application provides a method for evaluating the throughput of a semiconductor device, and the semiconductor structure provided by the embodiment of the present application will be described in detail below with reference to the accompanying drawings. FIG. 1 is a flowchart illustrating a method for evaluating the throughput of a semiconductor device according to an embodiment of the present application; FIG. 2 is another detailed flowchart of a method for evaluating the throughput of a semiconductor device according to an embodiment of the present application; fig. 3 is a flowchart illustrating a method for evaluating the throughput of a semiconductor device according to an embodiment of the present disclosure.
Referring to fig. 1 to 3, a method for evaluating the throughput of a semiconductor device including different types of process tools, the evaluating method comprising the steps of:
s101: and utilizing all the process machines to carry out production processing on the N batches of wafers, and acquiring data of all the process machines, wherein the data is the time interval between the production end of the current batch of the process machines and the production end of the previous batch of the current batch.
It should be noted that, due to the complex process flow and the dynamic and variable production line conditions, when all the process tools perform production processing on wafers, the number of wafers produced by a certain batch is not the maximum number of wafers produced by the certain batch in different process tools and/or different process steps, or when a certain batch of process tools process wafers, all chambers in the process tools are not in a working state, or the process tools of the current batch are down, so that the time interval between the end of the production of the current batch of process tools and the end of the production of the previous batch of the current batch of process tools becomes large, and so on. In the above situation, the obtained data of the characterization time interval of the process tool is not the data of the process tool in the full-load and steady-operation states, and these situations all affect the accuracy of the obtained data of the characterization time interval of all the process tools. In addition, the acquired data carries information indicating which batch the data is, which chambers of which process tools are in operation, which products are produced, and which process steps are performed.
S102: and acquiring effective data in the data as effective data.
The step of acquiring valid data in the data as valid data comprises: based on the maximum processing amount, the maximum processing amount is the maximum number of wafers allowed to be processed in any batch for the same type of process equipment, and the data corresponding to the maximum processing amount of the wafers processed in the current batch is reserved as valid data.
In this step, when the number of the wafers produced in the current batch is not the maximum number of the wafers that can be produced in the batch, the obtained data representing the time interval between the end of production of the current batch and the end of production of the next batch of the current batch is screened out, and the remaining data is kept as the valid data, so that the number of the wafers produced in the current batch represented by the valid data is ensured to be the maximum number, and the accuracy of the obtained valid data is improved.
It should be noted that the maximum number of wafers allowed to be processed by different types of processing tools in any one lot may be the same or different, and the embodiment of the present application does not limit the maximum number of wafers allowed to be processed by the processing tools in any one lot.
S103: at least two initial data sets are obtained, and data in each initial data set are effective data of each batch when the same type of process machine table produces the same product and carries out the same process step.
Because the process flow is complex and the production line condition is dynamically changeable, different products may have the same process steps when being manufactured, for example, two products are both manufactured through an etching process, and different products may need different types of process machines to be matched for manufacturing, therefore, the types of the effective data acquired in the step S102 are complicated, the difference between two random effective data may be large, the effective data are grouped through the step S103, so that the same type of process machines produce the same product and the effective data of each batch are in the same initial data group when the same process step is performed, the difference between the effective data in the same initial data group is favorably reduced, the effective data in each initial data group is conveniently processed and analyzed, and the accuracy of the result generated after the effective data are analyzed and evaluated subsequently is favorably improved.
S104: and acquiring a first data set based on the initial data set, wherein the data in the first data set is effective data corresponding to the working state of all the chambers in the process machine table in the initial data set.
In this step, when a certain batch of process machines processes wafers, and all chambers in the process machines are not in a working state, the effective data of the batch are screened out, and the remaining effective data are located in the first data group, so as to ensure that all chambers represented by the effective data in the first data group are in a working state in the process machines of the current batch, so that the effective data in the first data group are all obtained when the process machines are in a full-load state, and the accuracy of the retained effective data is further improved.
S105: and sorting the data in the first data group to form a first queue data group, and acquiring the standard deviation of the data in the first queue data group.
The sorting processing comprises ascending sorting or descending sorting, and therefore other processing on the effective data is convenient to perform subsequently. The standard deviation represents the dispersion degree of the effective data in the first queue data group, and the dispersion degree of the effective data in the first queue data group is larger if the standard deviation is larger, so that the obtained standard deviation can be used as a basis for judging whether the variation degree of the difference value of the adjacent front and back data in the first queue data group is too large or not.
S106: and acquiring a second queue data group based on the first queue data group and the standard deviation, wherein the second queue data group is formed by sorting the data left after removing the next data in the previous and next data when the difference value of the adjacent previous and next data in the first queue data group deviates from the standard deviation.
The step of acquiring the second queue data group comprises the following steps:
providing a first coefficient, wherein the deviation of the difference between adjacent pre-and post-data in the first queue data set from the standard deviation is represented by a first relation as follows:
|T NextLot -T PrevLot |>C 0 *σ
wherein, T PrevLot Representing the previous valid data, T, corresponding to the previous and subsequent data in the first queue data set NextLot Representing the next valid data corresponding to the previous and next data in the first queue data set, C 0 Denotes the first coefficient, σ denotes the standard deviation;
and removing the next effective data from the previous data and the next data which meet the first relational expression, and sequencing the remaining effective data to form a second queue data group.
The first relational expression is used for judging whether the variation degree between the adjacent front effective data and the adjacent back effective data in the first queue data group is within an allowable range, and when the variation degree between the adjacent front effective data and the adjacent back effective data is not within the allowable range, the difference between the next effective data and other effective data in the adjacent front effective data and the adjacent back effective data is larger, the effective data can be regarded as inaccurate effective data, the inaccurate effective data is screened out, for example, the data acquired when the process machine is in a downtime state is screened out, and the data acquired when the process machine is not in a stable operation due to other factors is screened out, so that the remaining effective data is located in the second queue data group, the accuracy of the effective data in the second queue data group is improved, the dispersion degree of the effective data in the second queue data group is reduced, and the accuracy of a result generated after the effective data is analyzed and evaluated subsequently is improved.
Wherein the first coefficient may range from 0.001 to 0.05. When the degree of dispersion of the valid data in the first queue data group is large, the first coefficient may be selected to be large, that is, the degree of variation between the valid data before and after being adjacent in the first queue data group is allowed to be large, for example, the first coefficient may be 0.049 or 0.05; when the degree of dispersion of the valid data in the first queue data group is small, the first coefficient may be selected to be smaller, that is, the degree of variation between the valid data before and after the first queue data group is allowed to be small, for example, the first coefficient may be 0.001 or 0.002.
In other embodiments, the following data corresponding to the deviation of the actual variation amplitude of the preceding and following data in the first queue data group from the average amplitude is represented by the following first relation:
|T NextLot -T PrevLot |≥C 0 *σ
s107: and acquiring minimum valid data and maximum valid data based on data in all the second queue data groups.
The method for acquiring the minimum effective data and the maximum effective data comprises the following three modes:
in some embodiments, referring to fig. 1, the least significant data and the most significant data may be directly obtained from all the second queue data groups obtained in step S106. In steps S101 to S106, the obtained data representing the time interval between the end of the production of the current batch and the end of the previous batch of secondary production of the current batch is screened for multiple times, so that the valid data in the second queue data set not only represents that the number of wafers produced by the current batch is the maximum number, but also represents that all chambers in the process tools of the current batch are in a working state, that is, it is ensured that the valid data in the second queue data set are all obtained when the process tools are in a full load state. Therefore, the accuracy of the effective data in the second queue data group is improved, so that the accuracy of the acquired minimum effective data and the acquired maximum effective data is improved, and the accuracy of a result generated after the effective data is analyzed and evaluated subsequently is improved.
In other embodiments, referring to FIG. 2, at step 106: after the second queue data set is acquired, at step 107: before the minimum valid data and the maximum valid data are acquired, the evaluation method can further comprise the following steps:
s110: and acquiring a first number and a second number, wherein the first number is the number of effective data in the second queue data group, and the second number is the number of effective data in the first data group corresponding to the second queue data group.
S120: and setting a first preset value and a second preset value according to the discrete degree of the effective data in the first data group.
S130: and reserving a second queue data set with the second number larger than the first preset value and the ratio of the first number to the second number larger than the second preset value.
S110 to S130, screening the plurality of acquired second queue data sets, and reserving second queue data sets with the second number larger than the first preset value, namely ensuring that the number of effective data in the first data sets corresponding to the second queue data sets is enough, so that the first data sets are processed subsequently to obtain second queue data sets, wherein the second queue data sets have reference values; further, a second queue data group with a ratio of the first number to the second number larger than a second preset value is reserved, that is, it is ensured that the second queue data group has a sufficient number of valid data relative to the first data group corresponding to the second queue data group. Therefore, steps S110 to S130 are beneficial to improve the value of the reserved second queue data set as a reference sample, so as to improve the accuracy of the minimum valid data and the maximum valid data obtained subsequently.
In still other embodiments, referring to fig. 3, in step S104: after the first data group is acquired, at step S105: before the sorting process is performed on the data in the first data group, the evaluation method may further include:
s140: acquiring a second number, wherein the second number is the number of effective data in the first data group; setting a first preset value according to the discrete degree of the effective data in the first data group; and reserving the first data group with the second number larger than the first preset value.
This step is advantageous to ensure that the number of valid data in the first data set is sufficient and suitable for use as a reference sample, so that the first data set is subsequently processed to obtain a second queue data set, which has a reference value. In addition, the first data group is screened before the step S105, which is beneficial to reducing the number of the first data group, so that when subsequent operations are performed according to the first data group, less effective data can be processed, time required for performing subsequent steps is reduced, and the evaluation efficiency of the evaluation method is improved.
In step S106: after acquiring the second queue data group, in step S107: before acquiring the minimum valid data and the maximum valid data, the evaluation method may further include:
s150: acquiring a first number, wherein the first number is the number of effective data in the second queue data group; setting a second preset value according to the discrete degree of the effective data in the first data group; and reserving a second queue data group of which the ratio of the first number to the second number is greater than a second preset value.
This step is advantageous to ensure that the second queue data group has a sufficient number of valid data relative to the first data group corresponding thereto, and since the number of valid data in the second queue data group is smaller than the number of valid data in the first data group corresponding thereto, the significance of the second queue data group as a reference sample is not large, which may adversely affect the accuracy of the minimum valid data and the maximum valid data to be subsequently obtained. Therefore, the part of the second queue data group is screened out, and the accuracy of the minimum effective data and the maximum effective data acquired subsequently can be improved.
It should be noted that, in the two embodiments of screening the second queue data sets, the step of obtaining the minimum valid data and the maximum valid data based on the data in all the second queue data sets includes: and acquiring minimum valid data and maximum valid data from all the remaining second queue data groups.
Moreover, the range of the first preset value may be 10 to 100, and the range of the second preset value is 10% to 80%, because the first preset value and the second preset value are both set according to the discrete degree of the valid data in the first data group, when the discrete degree of the valid data in the first data group is larger, the first preset value and the second preset value may both be selected to be smaller, for example, the first preset value may be 10 or 15, and the second preset value may be 10% or 15%; when the degree of dispersion of the valid data in the first data set is small, the first preset value and the second preset value can be both selected to be larger, for example, the first preset value can be 95 or 100, and the second preset value can be 75% or 80%.
In the above three embodiments, obtaining the minimum valid data and the maximum valid data includes the following two ways:
in some embodiments, obtaining the least significant data and the most significant data comprises the steps of:
and acquiring the maximum value and the minimum value of the effective data in each second queue data group, wherein the minimum effective data is the minimum effective data in the minimum values, and the maximum effective data is the maximum effective data in the maximum values.
In other embodiments, obtaining the least significant data and the most significant data comprises the steps of:
and acquiring a first number and a second number, wherein the first number is the number of effective data in the second queue data group, and the second number is the number of effective data in the first data group corresponding to the second queue data group.
And taking the second queue data group with the maximum ratio of the first number to the second number as a reference group, wherein the minimum valid data is the minimum value of valid data in the reference group, and the maximum valid data is the maximum value of the valid data in the reference group.
S108: and acquiring the production capacity of the semiconductor equipment based on the minimum effective data and the maximum effective data.
The method for acquiring the production capacity of the semiconductor equipment comprises the following steps:
providing a second coefficient and a third coefficient;
providing a maximum throughput, wherein the maximum throughput is the maximum number of wafers allowed to be processed in any batch by the same type of processing machine;
the semiconductor device throughput is expressed by the following second relational expression:
WPH∈[Run Size/(M 2 *Max)，Run Size/(M 1 *Min)]
wherein WPH represents a semiconductor device production capacity, run Size represents a maximum processing amount, and M 1 Denotes the second coefficient, M 2 Denotes a third coefficient, min denotes minimum significant data, and Max denotes maximum significant data.
Since the minimum valid data and the maximum valid data obtained in the above steps are highly accurate, the boundary values of the semiconductor device throughput and the ranges formed by the boundary values obtained by the second relational expression are also highly accurate.
The second coefficient may be in a range of 0.999 to 1, and the third coefficient may be in a range of 1 to 1.001. In the actual operation process of the semiconductor equipment, due to the existence of various factors (such as the fact that the semiconductor equipment is down or not in a full-load state), the acquired Takt Time fluctuates greatly along with the change of production batches, and the abnormal fluctuation data cannot reflect the normal level of the equipment. After the Takt Time is screened through the steps, the accuracy of the obtained minimum effective data and the maximum effective data is improved. In addition, the concept of adding the second coefficient and the third coefficient in step S108 is advantageous to further improve the accuracy of the boundary value of the finally obtained semiconductor device throughput and the range formed by the boundary value, thereby further improving the accuracy of the evaluated semiconductor device throughput.
In summary, by the above evaluation method, not only the acquired Takt Time is screened at least three times, but also the second queue data group composed of valid data is screened, which is beneficial to improving the accuracy of the acquired minimum valid data and maximum valid data, so as to improve the accuracy of the subsequently evaluated semiconductor device production capacity. In addition, the minimum effective data and the maximum effective data are processed subsequently to obtain the boundary value of the production capacity of the semiconductor equipment and the range formed by the boundary value, so that the accuracy of the evaluated production capacity of the semiconductor equipment is further improved, and the accuracy of the investment amount required by capacity expansion evaluation is improved.
Another embodiment of the present application further provides an apparatus for evaluating the throughput of a semiconductor device, which is used to implement the method for evaluating the throughput of a semiconductor device in the foregoing embodiments. An apparatus for evaluating the throughput of a semiconductor device according to another embodiment of the present application will be described in detail with reference to the accompanying drawings. Fig. 4 is a functional block diagram of an apparatus for evaluating throughput of a semiconductor device according to another embodiment of the present application.
Referring to fig. 4, an apparatus for evaluating the throughput of a semiconductor device includes: the data collection module 401 is configured to obtain data of all process tools, where the data is a time interval between the end of production of a current batch of the process tools and the end of production of a previous batch of secondary products of the current batch; a data processing module 402 for processing data, the data processing module 402 configured to: acquiring effective data in the data as effective data; acquiring at least two initial data sets, wherein the data in each initial data set is effective data of each batch when the same type of process machine table produces the same product and carries out the same process step; acquiring a first data group based on the initial data group, wherein the data in the first data group are effective data corresponding to the working state of all chambers in the process machine in the initial data group; sorting the data in the first data group to form a first queue data group; the second queue data group is obtained by sorting data left after removing the next data in the front and back data when the difference value of the adjacent front and back data in the first queue data group deviates from the standard deviation; an obtaining module 403, configured to obtain minimum valid data and maximum valid data based on data in all the second queue data sets, and obtain semiconductor device production capacity based on the minimum valid data and the maximum valid data.
Wherein, the data processing module 402 comprises: a validity filter unit 412, configured to obtain valid data in the data as valid data; a grouping unit 422, configured to obtain at least two initial data sets, where data in each initial data set is valid data of each batch when the same type of process machine produces the same product and performs the same process step, and obtain a first data set based on the initial data sets, where data in the first data set is valid data corresponding to all chambers in the process machines in the initial data set being in a working state; a sorting unit 432, configured to perform sorting processing on data in the first data group to form a first queue data group; a first calculating unit 442, configured to obtain an average amplitude corresponding to adjacent valid data in the second queue data group; the screening unit 452 is configured to obtain a second queue data group based on the first queue data group and the standard deviation, where the second queue data group is formed by sorting data left after removing the next data in the previous and next data when a difference value between adjacent previous and next data in the first queue data group deviates from the standard deviation, that is, after removing the next data in the previous and next data, sorting effective data left in the first queue data group to form the second queue data group.
In some embodiments, the data processing module 402 includes a first storage unit 414 and a second storage unit 424, wherein the validity filtering unit 412, the grouping unit 422, and the sorting unit 432 are all located in the first storage unit 414, and the first calculating unit 442 and the screening unit 452 are all located in the second storage unit 424. The data processing module 402 further includes: the data handling unit 462 is configured to transfer the valid data in the first queue data set to the first computing unit 442, i.e., transfer the data finally stored in the first storage unit 414 to the second storage unit 424.
In other embodiments, the validity filter unit, the grouping unit, the sorting unit, the first calculation unit and the screening unit are all located in the same storage unit, so that no data handling unit is required for data transmission between the sorting unit and the first calculation unit.
Wherein, the obtaining module 403 includes: a coefficient allocating unit 413, configured to provide the second coefficient, the third coefficient and a maximum throughput, where the maximum throughput is a maximum number of wafers that are allowed to be processed in any lot for the same type of processing tool; the second calculating unit 423 is used for calculating the throughput of the semiconductor device according to the minimum valid data, the maximum valid data, the second coefficient, the third coefficient and the maximum processing amount.
In summary, the apparatus for evaluating semiconductor device throughput can screen the acquired Takt Time at least three times, and also screen the second queue data group composed of valid data, which is beneficial to improving the accuracy of the acquired minimum valid data and maximum valid data, so as to improve the accuracy of the subsequently evaluated semiconductor device throughput. In addition, the minimum effective data and the maximum effective data are processed subsequently to obtain the boundary value of the production capacity of the semiconductor equipment and the range formed by the boundary value, so that the accuracy of the evaluated production capacity of the semiconductor equipment is further improved, and the accuracy of the investment amount required by capacity expansion evaluation is improved.
It will be understood by those of ordinary skill in the art that the foregoing embodiments are specific examples for carrying out the present application, and that various changes in form and details may be made therein without departing from the spirit and scope of the present application in practice. Various changes and modifications may be effected therein by one skilled in the art without departing from the spirit and scope of the application, and it is intended that the scope of the application be limited only by the claims appended hereto.
Claims (17)
1. A method for evaluating the throughput of a semiconductor device, the semiconductor device including different types of process tools, comprising:
utilizing all the process machines to carry out production processing on N batches of wafers, and acquiring data of all the process machines, wherein the data is a time interval between the end of the production of the current batch of the process machines and the end of the production of the previous batch of the current batch of the process machines;
acquiring effective data in the data as effective data;
acquiring at least two initial data sets, wherein the data in each initial data set is the effective data of each batch when the same type of process machine platform produces the same product and carries out the same process step;
acquiring a first data group based on the initial data group, wherein data in the first data group is the effective data corresponding to the working state of all chambers in the process machine table in the initial data group;
sorting the data in the first data group to form a first queue data group, and acquiring a standard deviation of the data in the first queue data group;
acquiring a second queue data group based on the first queue data group and the standard deviation, wherein the second queue data group is formed by sorting data left after removing the next data in the previous and next data when the difference value of the adjacent previous and next data in the first queue data group deviates from the standard deviation;
acquiring minimum effective data and maximum effective data based on data in all the second queue data groups;
and acquiring the production capacity of the semiconductor equipment based on the minimum effective data and the maximum effective data.
2. The method of claim 1, wherein the step of obtaining valid data of the data as the valid data comprises:
based on the maximum processing amount, which is the maximum number of wafers allowed to be processed in any batch for the process tools of the same type, the data corresponding to the maximum processing amount of the wafers processed in the current batch is reserved as the valid data.
3. The method of claim 1, wherein the step of obtaining the second queue data set comprises:
providing a first coefficient, wherein the deviation of the difference between adjacent pre-and post-data in the first queue data set from the standard deviation is represented by a first relation as follows:
|T Next Lot -T Prev Lot |＞C 0 *σ
wherein, T Prev Lot Representing the previous valid data, T, corresponding to the previous and subsequent data in the first queue data set Next Lot Representing the next valid data corresponding to the previous and next data in the first queue data group, C 0 Represents the first coefficient, σ represents the standard deviation;
and removing the next effective data in the front and back data corresponding to the first relational expression, and sequencing the remaining effective data to form the second queue data group.
4. The method of claim 3, wherein the first factor is in a range of 0.001 to 0.05.
5. The method of claim 1, wherein after acquiring the second queue data set, prior to acquiring the least significant data and the most significant data, further comprising:
acquiring a first number and a second number, wherein the first number is the number of the effective data in the second queue data group, and the second number is the number of the effective data in the first data group corresponding to the second queue data group;
setting a first preset value and a second preset value according to the discrete degree of the effective data in the first data group;
and reserving the second queue data group of which the second number is greater than the first preset value and the ratio of the first number to the second number is greater than the second preset value.
6. The method of claim 1, wherein after acquiring the first data set, prior to performing the sorting process on the data in the first data set, further comprising:
acquiring a second number, wherein the second number is the number of the effective data in the first data group;
setting a first preset value according to the discrete degree of the effective data in the first data group;
reserving the first data group with the second number larger than the first preset value;
after acquiring the second queue data group, the method further comprises:
acquiring a first number, wherein the first number is the number of the effective data in the second queue data group;
setting a second preset value according to the discrete degree of the effective data in the first data group;
and reserving the second queue data group of which the ratio of the first number to the second number is greater than the second preset value.
7. The method of claim 5 or 6, wherein the step of obtaining the least significant data and the most significant data based on data in all of the second queue data groups comprises:
and acquiring the minimum valid data and the maximum valid data from all the remaining second queue data groups.
8. The method of claim 7, wherein the first predetermined value is in a range of 10 to 100.
9. The method of claim 7, wherein the second preset value ranges from 10% to 80%.
10. The method of claim 1, wherein the step of obtaining the least significant data and the most significant data comprises:
and acquiring the maximum value and the minimum value of the effective data in each second queue data group, wherein the minimum effective data is the minimum effective data in the minimum values, and the maximum effective data is the maximum effective data in the maximum values.
11. The method of claim 1, wherein the step of obtaining the least significant data and the most significant data comprises:
acquiring a first number and a second number, wherein the first number is the number of the effective data in the second queue data group, and the second number is the number of the effective data in the first data group corresponding to the second queue data group;
and taking the second queue data group with the maximum ratio of the first number to the second number as a reference group, wherein the minimum valid data is the minimum value of the valid data in the reference group, and the maximum valid data is the maximum value of the valid data in the reference group.
12. The method of claim 10 or 11, wherein the step of obtaining the semiconductor device throughput based on the least significant data and the most significant data comprises:
providing a second coefficient and a third coefficient;
providing a maximum throughput, which is the maximum number of wafers allowed to be processed in any batch for the same type of the processing tools;
the semiconductor device throughput is expressed by a second relational expression as follows:
WPH∈[Run Size/(M 2 *Max)，Run Size/(M 1 *Min)]
wherein WPH represents the semiconductor device productivity, run Size represents the maximum processing amount, M 1 Represents the second coefficient, M 2 Represents the third coefficient, min represents the minimum significant data, and Max represents the maximum significant data.
13. The method of claim 12, wherein the second coefficient ranges from 0.999 to 1 and the third coefficient ranges from 1 to 1.001.
14. An apparatus for evaluating throughput of a semiconductor device, comprising:
the data collection module is used for acquiring data of all process machines, wherein the data is the time interval between the end of the production of the current batch of the process machines and the end of the production of the previous batch of the current batch;
a data processing module for processing the data, the data processing module configured to:
acquiring effective data in the data as effective data; acquiring at least two initial data sets, wherein the data in each initial data set is the effective data of each batch when the same type of process machine platform produces the same product and carries out the same process step; acquiring a first data group based on the initial data group, wherein data in the first data group is the effective data corresponding to the working state of all chambers in the process machine table in the initial data group; sorting the data in the first data group to form a first queue data group, and acquiring a standard deviation of the data in the first queue data group; acquiring a second queue data group based on the first queue data group and the standard deviation, wherein the second queue data group is formed by sorting data left after removing the next data in the previous and next data when the difference value of the adjacent previous and next data in the first queue data group deviates from the standard deviation;
and the acquisition module is used for acquiring minimum effective data and maximum effective data based on the data in all the second queue data groups and acquiring the production capacity of the semiconductor equipment based on the minimum effective data and the maximum effective data.
15. The apparatus of claim 14, wherein the data processing module comprises:
the validity filtering unit is used for acquiring valid data in the data as the valid data;
a grouping unit, configured to obtain at least two initial data sets, where data in each of the initial data sets is valid data of each batch when the same type of process machine produces the same product and performs the same process step, and obtain the first data set based on the initial data sets, where data in the first data set is the valid data corresponding to all chambers in the process machine in the initial data set being in a working state;
the sorting unit is used for sorting the data in the first data group to form a first queue data group;
a first calculating unit, configured to obtain a standard deviation of data in the first queue data group;
and the screening unit is used for acquiring the second queue data group based on the first queue data group and the standard deviation, wherein the second queue data group is formed by sorting data left after the next data in the previous and next data is removed when the difference value of the adjacent previous and next data in the first queue data group deviates from the standard deviation.
16. The apparatus of claim 15, wherein the data processing module further comprises:
and the data carrying unit is used for transferring the effective data in the first queue data group to the first computing unit.
17. The apparatus of claim 14 or 15, wherein the acquisition module comprises:
a coefficient configuration unit for providing a second coefficient, a third coefficient and a maximum throughput, wherein the maximum throughput is the maximum number of wafers allowed to be processed in any batch by the same type of the processing machine
A second calculation unit for calculating the semiconductor device throughput according to the minimum effective data, the maximum effective data, the second coefficient, the third coefficient, and the maximum throughput.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110924495.7A CN115705543A (en) | 2021-08-12 | 2021-08-12 | Method and apparatus for evaluating throughput of semiconductor device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110924495.7A CN115705543A (en) | 2021-08-12 | 2021-08-12 | Method and apparatus for evaluating throughput of semiconductor device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115705543A true CN115705543A (en) | 2023-02-17 |
Family
ID=85180883
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110924495.7A Pending CN115705543A (en) | 2021-08-12 | 2021-08-12 | Method and apparatus for evaluating throughput of semiconductor device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115705543A (en) |
-
2021
- 2021-08-12 CN CN202110924495.7A patent/CN115705543A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP3957964B2 (en) | Lot dispatching method and lot dispatching system | |
| CN101290611A (en) | Data abnormal point detection method and device | |
| CN110928265B (en) | Data processing method, device and system, and computer readable recording medium | |
| CN104701211B (en) | Inspect the method for measurement of frequency by random samples according to integrated circuit manufacture process Capability index adjust automatically | |
| US6821792B1 (en) | Method and apparatus for determining a sampling plan based on process and equipment state information | |
| CN108171435B (en) | Production plan decision method considering preventive maintenance | |
| CN115705543A (en) | Method and apparatus for evaluating throughput of semiconductor device | |
| CN115705544A (en) | Method and apparatus for evaluating throughput of semiconductor device | |
| CN105448817A (en) | Method for electrochemical polishing of metal interconnection wafer structure | |
| CN115705545A (en) | Method and apparatus for evaluating throughput of semiconductor device | |
| CN113408915B (en) | Production line bottleneck identification method and system | |
| CN110705844A (en) | Robust optimization method of job shop scheduling scheme based on non-forced idle time | |
| US7778798B2 (en) | System and method for measuring tool performance | |
| CN113850289B (en) | Working procedure working hour quota measuring and calculating method and system for radio frequency microwave product | |
| CN114384872A (en) | Product development process quality comprehensive management and control system | |
| CN106485401A (en) | Meet the enumeration value working procedure ability estimation method of Poisson distribution | |
| EP3460732B1 (en) | Dispatching method and system based on multiple levels of steady state production rate in working benches | |
| CN110426999A (en) | Statistical process control method and its control system | |
| CN108268987B (en) | Method for estimating quality of various products | |
| CN117609710B (en) | Method and device for preventing normal jump of monitoring data from being removed | |
| CN117708615B (en) | Product detection method and processing equipment monitoring method | |
| CN111755079B (en) | Method and system for determining optimal raw material proportioning scheme of polycrystalline silicon | |
| CN116230595B (en) | Device and method for monitoring comprehensive efficiency of semiconductor equipment | |
| US8332416B2 (en) | Specification establishing method for controlling semiconductor process | |
| US20020052667A1 (en) | Determining the capacity components of multi chamber systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |