JP2010015236A - Production line evaluation method, production line evaluation system, production line evaluation program, recording medium recorded with production line evaluation program, and production line control method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a production line evaluation method capable of obtaining the optimal production path on the basis of the performance of a product and the processing information of production. <P>SOLUTION: In the method for evaluating the production path of a production line for producing a product by a plurality of production paths, processing information to be operated for producing a product on each production path and inspection information to be acquired by inspecting the product to be produced by each production path are collected, the collected processing information and inspection information are made correspond to each other to acquire correspondence information. The plurality of production paths are ranked based on the correspondence information, and the ranked production paths are displayed in a display device. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a production line evaluation method for producing products by a plurality of production paths, a production line evaluation system, a production line evaluation program, a recording medium on which the production line evaluation program is recorded, and a production line control method. Furthermore, the present invention relates to a production line evaluation method, a production line evaluation system, and a production line control method for producing a plurality of product models through a plurality of production paths.

For example, production lines such as semiconductor devices take time from wafer loading to device completion, and there are many causes such as changes in manufacturing conditions, equipment maintenance, equipment failures, and the effects of fluctuations in yield and equipment availability. This reduces the production volume. In order to cope with the decline in production volume, production is carried out with a large amount of work in progress. However, increasing the in-process increases costs and prolongs production time. In addition, it has become an important issue to produce necessary products in a timely and lossless manner from the viewpoint of subdividing customer requirements and observing delivery dates. Such a problem occurs not only in the semiconductor field, but also in a manufacturing process having a plurality of processing steps and a plurality of inspection steps such as a steel plant and a chemical plant.
In particular, the production process of semiconductor devices has a very large number of processes, and the same product is different in the production process, but it passes through the same production equipment many times, and multiple products are produced by different production paths. The model is mixed and produced. Therefore, it is complicated and difficult to manage the in-process amount of each process. In the process of manufacturing a plurality of product models, there is a concern that bottlenecks may occur due to device maintenance, device failure, and the like. The challenge in the production process is that one problem always keeps answering customer requests without affecting other models.

Patent document 1 controls the production amount of each process according to the difference between the target value and the actual value of the in-process amount set for each process of the production process based on the production plan, and the in-process amount of the entire production process We have proposed a method for producing products with a value below the proper value. That is, the cumulative value of each process is calculated in reverse order from the final process, and the production amount of each process is controlled in accordance with the difference between the accumulated in-process target value and the actual value.
In addition, Patent Document 2 discloses a yield rate in order of increasing test rate at which the number of abnormal lots protrudes for each manufacturing device used in each process for each manufacturing lot and for each position in the manufacturing device for failure factor analysis in the production process. A method to identify the factor is proposed. That is, the yield factor is specified by associating the number of normal lots with the number of abnormal lots, or associating the average process time from the previous process end time to the process start time with the normal lot and the abnormal lot.
JP 2003-233410 A JP 2005-12095 A

  For example, in the production process of semiconductor devices, processing shape such as line width and oxide film thickness at the time of product processing, processing temperature, processing elapsed time, processing liquid, processing gas, processing voltage, voltage, current, processing speed, etc. A large amount of data such as electrical characteristics and yield is collected. Analysis of process data is performed with one of the main objectives being to extract defect factors of products using these data. For example, FIG. 1A is a scatter diagram showing the relationship between the product performance Q of a certain product and the process data X in the manufacturing process of the product. As shown in FIG. 1A, the product performance Q and the process data X show a strong correlation. FIG.1 (b) is a figure which distinguishes and represents the scatter diagram of Fig.1 (a) for every manufacturing apparatus (1st machine, 2nd machine, 3rd machine) used by the manufacturing process. From FIG. 1 (b), it can be seen that there are many defects in which the product performance Q deviates from the product performance management standard R in the products processed by the second machine. Therefore, when the product performance Q is to be controlled within the product performance management standard R, it is understood that it is most effective to take measures against the second machine that is the cause of the failure.

  However, it can be seen from FIG. 1C that this is not sufficient. In other words, FIG. 1C shows a scatter diagram when the model A and the model B produced by different production lines are processed under the same processing conditions in the same process. From FIG. 1 (c), it can be seen that, with model A and model B produced by different production lines, only model A deviates from product performance management standard R even when processed by the same second machine. Thus, when a plurality of product models are produced by the same manufacturing apparatus, it is necessary to perform data analysis in consideration of the model, production process, manufacturing apparatus, and the like. In other words, a single analysis or countermeasure for a problem is not sufficient, and in some cases, the problem can be prolonged. When performing production control in the production process, if unnecessary production control is performed, the entire manufacturing process may be confused, and defective products may continue to be generated, resulting in increased losses and supply defects.

  With respect to such a problem, Patent Document 1 is a method for producing by suppressing the in-process amount in the production process to an appropriate value, and Patent Document 2 is a method for identifying an abnormal lot and identifying a yield factor. However, production processing information such as the correlation between product performance and process data, and differences in manufacturing equipment or product models used in the manufacturing process are not considered. In addition, no production line evaluation method has been proposed that satisfies customer requirements and maximizes the number of products shipped or minimizes production loss.

Accordingly, a first problem of the present invention is to evaluate each production path based on processing information for production and product inspection information in a production line that produces products by a plurality of production paths. Further, it is to evaluate a plurality of production paths for a plurality of product models, respectively, and to evaluate a production path that is optimal as a whole so that production can be performed for all product models. The second problem is to minimize production loss while satisfying customer requirements.
Accordingly, the present invention provides a production line evaluation method including means for evaluating a production route for a single model. In addition, by evaluating the production path considering multiple product models, the production path in the entire production line can be quickly evaluated, and the difference between the predicted value and the actual value that is effective when using the evaluated production path is calculated. It is intended to provide a production line evaluation method for reducing the size. This facilitates the decision making of which model should be preferentially processed and which apparatus in which process.

  Another object of the present invention is to provide a production line evaluation system that executes such a production line evaluation method and helps to improve the yield in the manufacturing process. Another object of the present invention is to provide a production line evaluation program for causing a computer to execute the production line evaluation method. Moreover, this invention is providing the computer-readable recording medium which recorded the production line evaluation program. Furthermore, it is providing the production line control method which controls a production line according to the production path evaluated by the production evaluation method.

  In order to solve the above problems, a production line evaluation method of the present invention is an evaluation method for evaluating the productivity of a plurality of production paths in a production line having a plurality of production paths, and produces products through the production paths. A processing information selection step for selecting from a processing information storage unit for storing processing information at the time of processing, an inspection information collection step for collecting from an inspection information storage unit for storing inspection information of products produced by the production route, and the collection Linking the processed information and the inspection information to obtain linking information; and a ranking step of ranking the plurality of production paths in the order of productivity based on the linking information; A production route display step for displaying the ranked production routes on a display device.

  As shown in FIG. 2 (a), the production line evaluation method according to the present invention includes a plurality of manufacturing apparatuses such as No. 1 and No. 2 in process 1 and multi-stage production including one inspection apparatus in process 3. This is intended to evaluate the production path when selecting the first or second manufacturing apparatus in Step 1 for the line. Note that the process 2 includes only one manufacturing apparatus. The production line may be the entire process of the semiconductor manufacturing apparatus, or may be a part of the manufacturing process of the semiconductor manufacturing apparatus.

Further, as shown in FIG. 2B, the production line targeted by the present invention includes, for example, a first production line that produces model A, a second production line that produces model B, and a third that produces model C. It has multiple production lines like a line. Each of the first production line to the third production line has five processes, and each of the processes 1 to 4 has a manufacturing apparatus, and each manufacturing apparatus performs a predetermined process to perform multi-stage production. Produce products in line. Step 5 is an inspection device. Steps 1 and 4 include Units 1 to 3, and Steps 2 and 3 include Units 1 and 2. The production line shown in FIG. 2B is an exemplification, and each manufacturing apparatus and one inspection apparatus provided in Steps 1 to 4 may be the same or different. The inspection apparatus may include at least one inspection apparatus for three production lines. Note that the first production line to the third production line include a plurality of manufacturing apparatuses in at least one process.
Here, the “product” includes not only a finished product but also a work in progress such as a semi-finished product.

The processing information collected by the present invention is processing information applied to a semiconductor device. For example, a film forming device such as a semiconductor film, an insulating film, or a conductor film, a peeling device and an etching device for peeling each of the above films , Processing information such as a cleaning device and a drying device. Further, it may be waiting time information and repetition number information between each process.
The inspection information includes device performance such as electrical characteristics such as voltage and current of the semiconductor device, inspection results such as current consumption, processing speed, and durability, non-defective product rate, defect rate, and yield of the wafer test.
In each process of each production line, a semiconductor device is processed in units of production lots by a manufacturing apparatus. Further, the semiconductor device is inspected for each production lot by the inspection device. Here, the processing information performed by each manufacturing apparatus and the inspection information of each inspection apparatus are stored in association with each production lot. This association is referred to as linking.

Further, in the present invention, the ranked production route is a production route that maximizes the number of shipments calculated based on the production capacity for each production route, the number of shipments for each model is maximized, and the number of shipments of all models is Production path that maximizes total, minimizes production loss for each model, minimizes production loss for all models, priority for each model, scheduled shipping date or planned shipment quantity, and production for each production process Production path that minimizes loss, difference in delivery schedule for each model, production path that minimizes delivery schedule difference for all models, priority for each model, estimated delivery schedule, scheduled shipment quantity, and production capacity Based on the production route that minimizes the delivery schedule difference, inspection information and processing information based on the simulation information by the multiple regression analysis method, the production route extracted in order from the important items with the greatest impact, updated in a certain period, Constantly monitor the actual value and the predicted value in the case of using the manufacturing apparatus caul was steps a productive pathway with a correction process of reducing the variations.
Therefore, according to the above configuration of the present invention, a plurality of production paths can be evaluated based on product performance and processing information. In addition, by evaluating a plurality of production paths by computer processing as described above, it is possible to evaluate the production path before actually changing the production path, so it is possible to simulate an actual production line. .

Further, the production line evaluation method of the present invention ranks the plurality of production paths based on the analysis result, a production path analysis step for analyzing the plurality of production paths based on the association information. The analysis result ranking step and the production route display step are repeated to rank a plurality of production routes for a plurality of product models. With this configuration, when a plurality of types of products are produced through a plurality of production routes, the production routes for producing all the types of products are evaluated, thereby enabling the entire optimum production.
Furthermore, the production line evaluation method of the present invention includes an input unit for product plan information including a priority for each model of the product, a planned shipping schedule or a planned shipping quantity, and the ranking step includes the linking information and the It is characterized by ranking production routes based on product planning information. With this configuration, production loss can be minimized while satisfying customer requirements.

  In addition, the production line evaluation method of the present invention has the display means, displays the predicted value and actual value for each production route for every model, model, and production route, and grasps the production control status in real time. It is characterized by enabling verification.

According to another aspect, the present invention is a production line control method comprising a CIM for managing a production line, wherein the production path is controlled by the CIM according to the ranked production path. To do.
According to another aspect, the present invention is a program for causing a computer to execute the above steps. According to the program of the present invention, the computer can execute the production line evaluation method.
According to another aspect, the present invention is a computer-readable recording medium on which the program is recorded. According to the recording medium of the present invention, it is possible to cause the computer to execute the production line evaluation method by causing the computer to read the recorded content of the recording medium.

Furthermore, according to another aspect of the present invention, there is provided a production line evaluation system, wherein a plurality of production paths provided in the production line and processing information for storing processing information when products are produced through the production paths are stored. A processing information selection unit that is selected from the storage unit, an inspection information collection unit that is collected from an inspection information storage unit that stores the inspection information of products produced through the production path, and the collected processing information and the inspection information are linked. A linking unit that obtains linking information, a ranking unit that analyzes and ranks the plurality of production paths based on the linking information, and displays the ranked production paths on a display device. A production route display unit.

  According to the present invention, each production path can be evaluated based on processing information for production and product inspection information in a production line that produces products by a plurality of production paths. In addition, it is possible to evaluate a plurality of production paths for a plurality of product models, and to evaluate a production control path that is optimal as a whole. It is also possible to evaluate a production path that minimizes production loss while satisfying customer requirements.

The production line evaluation method of the present invention targets a multi-stage production line that produces a product by continuously performing at least one production process including a plurality of manufacturing apparatuses and an inspection process including an inspection apparatus. In this production line, in each process of the production line, for each product applied by the manufacturing apparatus, processing applied to produce the product is collected (processing information collection step). Further, processing information is collected in the order of processes (processing history information collecting step). Moreover, the inspection information showing the inspection result obtained in the inspection process including the inspection device is acquired (inspection information collecting step). Then, the processing information or the processing history information and the examination information are associated with each other and acquired as linking data (information linking step). When linking, it is more preferable to associate with each production lot of the product. Next, an analysis by a predetermined analysis method is performed using the association data, and the degree of influence on the inspection information (here, product performance) is analyzed for each of a plurality of manufacturing apparatuses (optimum production path analysis step). . As the predetermined analysis method, a multiple regression analysis method is preferable. The production path, which is the analysis result here, indicates which manufacturing apparatus of a certain process is used to produce the most influence on the inspection information in a certain model. Thereby, better inspection information can be expected by using the assigned manufacturing apparatus. Further, ranking is performed in the order of the best influence on the inspection information based on the analysis result of the optimum production path analyzing step (ranking step). The ranked production route information is displayed on a predetermined display screen (production route display step).
In the second embodiment of the present invention, in the ranking step, the priority of each model, the shipping schedule and the shipping schedule quantity input from the model plan input step are used for the analysis.

  Based on the ranked production route information, a user (a person who executes this production line evaluation method, specifically a person in charge of production line control management) estimates the optimum production in the process having the plurality of manufacturing apparatuses. It is possible to grasp the manufacturing apparatus to be used and to instruct the product flow. Alternatively, the optimum production path information is fed back to the CIM that manages the production line, and control is performed so that production is automatically performed through the optimum production path. Thus, a production line with only one model can be optimized by a single analysis through a series of steps.

  In addition, when producing a plurality of product models, the present invention repeatedly performs a production route analysis step, an analysis result ranking step, and a production route display step, and a plurality of product types used across a plurality of product models. Analyze and rank manufacturing equipment comprehensively. As a result, the production path that is optimal as a whole is evaluated so that a plurality of product models can be produced by the optimum production path. In other words, in a production line where multiple product models are mixed, not only one product model but also other product models are spread, and the production path is fluidized so that all product models can be optimally produced as a whole. Is. As a result, when a plurality of manufacturing devices are arranged in a certain process, and it turns out that a specific manufacturing device has a failure factor in a specific product model, the corresponding model does not use that specific manufacturing device, etc. Countermeasures are determined. Then, the production path is changed so that the production of the specific product model whose failure factor is found is produced by another manufacturing apparatus. As the production path of a specific product model is changed in this way, when the production equipment is used in common with other product models, the production path is redistributed so that the production quantity of the other product model is also achieved. There is a need. As a result of the redistribution, it can be confirmed by simulation whether or not it is possible to satisfy the production number, delivery date, and production loss of all product models by the optimum production path. The production route information that is obtained as a whole optimum in this way is controlled by the user or by the CIM.

Hereinafter, the best mode for carrying out the present invention will be described.
First, the analysis target of the optimal production path analysis used in the production line evaluation method of the present invention will be described. In the following, a clearing path evaluation method for a production line that produces semiconductor devices will be described. However, the production line evaluation method of the present invention is a manufacturing process having a plurality of processing steps and an inspection step, such as a steel plant or a chemical plant. Is also applicable.
The inspection result information indicating the product performance of the product includes, for example, a non-defective product rate or a defect rate and a yield of a wafer test for a finished product or a semi-finished product such as a semiconductor device in the semiconductor device manufacturing field. In addition, device electrical property tests such as voltage and current, current consumption, processing speed, durability, and the like for a semiconductor device finished product or semi-finished product using a wafer test are included. Furthermore, processed shapes (oxide film thickness, line width, overlay value, foreign matter inspection value, pattern defect value) measured in each manufacturing process of the product or semi-finished product can also be used. Using the inspection result information as described above, the optimum production path that maximizes the number of models that satisfy the product performance is analyzed (manufacturing quality perspective). In addition, the inspection information is weighted by using the model priority determined from the customer request, the scheduled shipping schedule, the scheduled shipping quantity, and the production capacity of each production process on the production line. As a result, the production loss is calculated, and the optimum production path that minimizes the production loss in all models is analyzed (from the viewpoint of production technology). In addition, by weighting the inspection information, a delivery schedule difference is calculated, and an optimal production path that minimizes the delivery schedule difference in all models is analyzed (production management perspective).

  Further, among the production paths evaluated by the production line evaluation method, which production path is used (which manufacturing apparatus is used in which manufacturing process) is a simulation result based on the ranked production path. I want to know before instructing product flow. In such a case, the user designates, for example, the highest-order production path (process and manufacturing apparatus) from the above-described ranked production paths, so that the manufacturing apparatuses in the remaining other processes are analyzed and ranked again. It can be done instantly and presented to the user. In this case, the user can quickly grasp the effect produced by the optimum production route specified based on the production line evaluation result information in advance as a predicted value, and therefore can easily make a decision on product flow. Can do.

  According to the production line evaluation method of the present invention, it is possible to grasp the current optimum production path in the entire production process in a plurality of types of production lines having a plurality of manufacturing apparatuses in at least one process. By comparing the difference between the predicted value and the actual value when using, the bias in the entire production process can be eliminated. Therefore, the production process can be optimized by performing the production using the production apparatus that is optimal in the production process, and the yield can be improved.

(First embodiment)
FIG. 3 is a diagram for explaining the overall configuration of a production apparatus provided with the production line evaluation system of the present invention.
The production apparatus 1 uses a CIM (Computer Integrated Manufacturing) system 2 for managing the manufacturing process, a customer information management system 3 for managing customer information, a storage medium storing the production line evaluation program of the present invention, and the program. A production line evaluation system 4 for causing a computer to execute the production line evaluation method of the present invention is provided, and these systems are connected to each other via a data communication path 5. Here, the data communication path 5 may be wired or wireless, and is connected by a LAN (Local Area Network).

  The CIM system 2 manages a multi-stage production line in which at least one production process including a plurality of manufacturing apparatuses and an inspection process having an inspection apparatus are made continuous. Alternatively, a plurality of production lines having a plurality of manufacturing apparatuses in each process of the plurality of production lines are managed. That is, a production line as shown in FIG. 2A or 2B is managed. Further, the CIM system 2 has a control function for selecting or changing a production route according to production route information fed back from the production line evaluation system 4. At the same time, it includes a database 2a that stores processing information of each manufacturing apparatus and inspection result information of the inspection apparatus. More preferably, the processing information is collected in the order of processes and used as processing history information.

The processing information includes processing information such as a film forming apparatus such as a semiconductor film, an insulating film, a conductor film, and a photosensitive film, a peeling apparatus and an etching apparatus that peels each of the above films, a patterning apparatus, a cleaning apparatus, and a drying apparatus. Including. For example, in a film forming apparatus such as MBE, the temperature, pressure, gas type, concentration, processing profile, acceleration voltage, and the like during film formation. In a peeling apparatus such as a plasma etching apparatus, for example, the type of gas, pressure, temperature, voltage, time, and the like. In the etching apparatus, for example, the type, concentration, temperature, processing time, etc. of the processing liquid. In the cleaning apparatus, for example, the cleaning liquid, the cleaning time, the number of repetitions, and the like. In the drying apparatus, the drying temperature, the drying time, and the like.
The processing information includes processing shape information such as line width, oxide film thickness, overlay value, foreign matter inspection value, pattern defect value, and processing device name, processing temperature, processing elapsed time, processing liquid, and the like when processing the semiconductor device. Processing gas, processing voltage, processing date and time, product name, model name, production lot number, processing person name are included. Furthermore, a processing chamber and a processing position in a furnace such as a diffusion furnace included in each processing apparatus are included. Furthermore, you may include the waiting time between each process, and the repetition frequency of each process.

  The inspection information includes the inspection device name, product name, model name, production lot number, inspection conditions, device performance such as electrical characteristics such as voltage and current, current consumption, processing speed, and durability. , Wafer test non-defective rate, defect rate and yield, inspection date and time, inspector name, pass / fail judgment result, product name, production lot. These can be collected by a semiconductor wafer test apparatus.

  The customer information management system 3 manages the customer name, product name, model name, order date, number of orders, priority order, shipping date, number of shipments, etc. for each customer. In order to manage these, an input unit (not shown) is provided, and from this input unit, customer request information including the product model name, priority for each product model, planned shipping schedule, and planned shipping quantity, which is a request from the customer. A database 3a to be stored is included.

  The production line evaluation system 4 acquires processing information and inspection information from the CIM system 2 that manages the manufacturing process via the data communication path 5. Further, customer request information for each product model is acquired from the customer information management system 3 as necessary. And the production line evaluation system 4a which analyzes the order of the optimal production path based on these information is provided. The optimum production path information analyzed by the optimum production line evaluation system 4a is fed back to the CIM system 2 to realize production line control by product flow that becomes the optimum production path. In order for the computer to realize the above operation, a production line evaluation program is stored in the database 4b.

  In this way, the production line evaluation system 4 automatically calculates the optimum production path in consideration of all product models according to the program. The “production control step” for reflecting the calculated and calculated optimum production path in the production process is fed back to the CIM system that manages the production process when it is reflected by a person (production control manager). It may be executed automatically. In this way, the number of products shipped is maximized by configuring the product flow to be optimized as a whole. Also satisfy customer requested plan. It also minimizes production loss during the production process. As described above, productivity can be improved and product performance can be improved.

FIG. 4 shows the configuration of a production line evaluation system 4 provided with the production line evaluation method of the present invention.
The production line evaluation system 4 of the present invention includes an association data table 11 as a storage unit, an optimum production route analysis template unit (framework) 12 for performing optimum production route analysis, a variable selection operation unit 13, and an input processing unit. The optimum production route analysis operation unit 14 and a display device 15 having a display screen are provided.

The variable selection operation unit 13 is mainly used by the user to specify an analysis target from the inspection result information and processing information or processing history information of each product stored in the association data table 11 shown in FIG. . The variable selection operation unit 13 inputs characters and symbols in the same manner as an ordinary computer input device, and gives instructions and instructions to the computer. Accordingly, the variable selection operation unit 13 includes a keyboard, a mouse, or a pointer, and the production control system 4 has software for processing input information.
The optimum production path analysis operation unit 14 designates inspection result information for displaying an analysis result among a plurality of types of inspection result information on a display screen shown in FIG. Used to specify. Therefore, the optimum production path analysis operation unit 14 is also composed of a keyboard, a mouse, or a pointer, and is used in common with the variable selection operation unit 13.

  The optimal production route analysis template unit 12 includes a data editing function unit 12a serving as a linking processing unit, a multiple regression analysis function and ranking unit 12b serving as a ranking unit, and a process step designating unit for designating an optimal production route. 12c, except for the designated processing step, a ranking and repeating unit 12d for performing new ranking in other steps, and a graph display function unit 12e for performing display processing.

The data editing function unit 12a outputs M pieces of inspection result information 11a and N pieces of processing history information 11b to be subjected to the multiple regression analysis stored in the association data table shown in FIG. Edit the link to correspond to each.
The multiple regression analysis function and ranking unit 12b is edited by the data editing function unit 12a, and N pieces of processing history information 11b associated with each production lot of the product are used as explanatory variables, and M pieces of inspection result information 11a. Is used as the objective variable. The objective variable is M pieces of inspection result information 11a, and indicates the number of products that satisfy the shipping determination criteria. Here, the number of products is the total number of products that have passed an inspection such as a device electrical characteristic test.
In other words, the number of products is the total number of products that satisfy all the shipping determination conditions in the inspection result indicating the product performance in the production process. That is, as shown in Table 1, the number of products satisfying the shipping determination condition is aggregated as the number of shipments for each product lot, and the number of shipments for each model, that is, the total number of shipments for each product lot is calculated. Further, the total number of shipments for each model is calculated to calculate the number of shipments for the entire product.

上記説明変数と目的変数を用いて重回帰分析を行い、出荷数に影響を与える分散比（Ｆ値）を求め、これを大きい順に並べ処理工程名で表示する。

Multiple regression analysis is performed using the explanatory variable and the objective variable, and a variance ratio (F value) that affects the number of shipments is obtained, and these are arranged in descending order and displayed by process step names.

As described above, the processing step designating unit 12c designates the processing steps ranked by the multiple regression analysis function and the ranking unit 12b in order from the processing step having a particularly large influence on the inspection result information.
As described above, the ranking and repetition unit 12d is configured by the multiple regression analysis function and the ranking unit 12b, except for the processing steps that have a large influence on the inspection result information, among other processing steps. Make a new ranking in the process. In this manner, except for processing steps having a large influence on the inspection result information, new ranking operations in other steps are repeatedly performed. This reduces product variation.
The graph display function unit 12e displays the relationship between the inspection result information ranked by the ranking and repeating unit 12d and the processing history information as described in FIG.

  The user causes the graph display unit 12e to draw a graph of the relationship between the ranked inspection result information and the processing history information, and the production route analysis operation unit 14 instructs the ranking repeating unit 12d to perform repeated analysis. By this repeated analysis, a graph is drawn for each ranked processing step, so that the predicted value and the actual value can be grasped, and the processing steps as the optimum production path can be narrowed down more easily.

FIG. 5 shows a flowchart for explaining the production line evaluation method of the present invention including the optimum production route analysis. The flow chart of FIG. 5 uses, for example, the non-defective product rate, the defective rate, and the yield of the wafer test as the first analysis target in the production process of the semiconductor manufacturing field, and uses the inspection result information indicating these product performances, It is a flowchart which maximizes the number of products shipped satisfying the shipping determination condition.
The number of products shipped that is the first analysis target is the total number of products that satisfy all the shipping judgment conditions in the inspection result indicating the product performance in the production process. That is, as shown in Table 1.

  The optimum production path that maximizes the number of products shipped for each product model is the production path that maximizes the number of products that meet the shipping standards in the production process, that is, the path that is produced by a highly productive device in each manufacturing process. . Optimization of the entire product production process means that the effective schedule for each process is selected by prioritizing the manufacturing equipment with limited production capacity while maintaining the expected shipping schedule and the planned shipping quantity as customer requirements. This is a production route that maximizes the number of shipments for all models.

The flowchart of FIG. 5 shows a processing procedure of a production line evaluation method for a multi-stage production line that has a plurality of manufacturing apparatuses in at least one process and produces a plurality of types of products. The production line covers a manufacturing process including at least one inspection process for inspecting product performance of a product. Therefore, each production line does not have to have an inspection device.
The first step S1 of this flowchart is an inspection information collection step for collecting the inspection results of the product performance by the inspection device. The next step S2 is a processing information collection step for storing the processing performed by each manufacturing apparatus on the production line and collecting product processing information. More preferably, the processing information is stored in the order of the processes and used as processing history information. The order of steps S1 and S2 may be reversed or may be processed simultaneously. Next, step S3 is an information linking step for linking the inspection result information and the processing history information in association with the production lot. Next, step S4 is a production route analysis step of analyzing the production route by multiple regression analysis from the pegging information. Thus, according to the analyzed production path, step S5 is a ranking step for ranking according to the degree of influence on product performance. Step S6 is a production control display step for displaying the production flow based on the production paths ranked in step S5.

By repeating the above-described series of steps (S4 to S6), the ranking step S5 is performed according to the degree of influence on product performance in a comprehensive manner across a plurality of manufacturing apparatuses for a plurality of product models. To rank. As a result, it is possible to carry out production through a production path that is optimal for all product models. The user can select the optimum production route ranked by the above flow. In this way, it is possible to select the production path in the order in which the inspection information is most effective rather than selecting the optimum production path. Each time the optimum production route is selected, the graph display unit 12e checks the difference between the predicted value and the actual value calculated by the optimum production route in the production control display step, and repeats the steps of analyzing and ranking again. Can do.
As a result, the user makes a decision on the product flow to the production process (S8) by the optimum production path that can be judged to be most effective. Here, although the user makes a decision, the CIM may automatically select it by feeding back to the CIM system 2 that manages the production process according to the instruction of the optimum production path.

FIG. 6 is a diagram for explaining the data table 11 created in the information linking step (S3).
The data table of FIG. 6 stores the inspection result information (11a) and the processing information (11b) for each product lot number (11c) for the target product or semi-finished product. Data obtained by collecting processing information in the order of processes is referred to as processing history information.
The inspection result information (11a) stored in the horizontal column of the data table is product performance of performance 1 to performance M. For example, in the semiconductor manufacturing field, the inspection result information includes the inspection device name, inspection conditions, device performance such as electrical characteristics, inspection results such as current consumption, processing speed, durability, wafer test non-defective rate, defect rate and yield, Includes inspection date / time, inspector name, pass / fail judgment result, product name, production lot. As processing history information (11b), processing shape information such as line width at processing, oxide film thickness, overlay value, foreign matter inspection value, pattern defect value, processing device name, processing temperature, processing elapsed time, processing liquid, Includes processing gas, processing voltage, processing date and time, product name, production lot, and name of person in charge of processing. In addition, a processing chamber and a processing position in a furnace such as a diffusion furnace included in each processing apparatus are included. Furthermore, the waiting time between each process is included.

  In the column of the data table of FIG. 6, the production lot number (11c) of the product is described as a common identifier, for example, a production lot number used for product identification in the semiconductor manufacturing field. With this data table, the inspection result information (11a) and the processing history information (11b) are associated with each other using the production lot number (11c) as a key, whereby the association data table 11 is obtained. In this association data table 11, the processing device name and the processing date and time of each process in the L identifiers (lot numbers) are stored in association (linked) with each other for each line. Further, in this data table, model names (11d) having different production lines are also stored as the association data. The association data table 5 includes not only the processing date and time of each process, but also the processing conditions such as processing temperature, processing time, type / concentration of processing liquid, processing pressure, processing voltage, inspection voltage, inspection current, inspection device, etc. Various data groups such as the inspection conditions can be stored. The data table shown in FIG. 6 is used for production path analysis in step S4.

FIG. 7A is a diagram for explaining a layout example of the optimum production route analysis template 30 used in the optimum production route analysis template unit 12 described with reference to FIG.
The optimum production route analysis template 30 is laid out so as to have a ranking result display unit 31, a graph display unit 32, and an analysis operation unit 33.
The ranking result display unit 31 refers to the linkage data table edited so that the M pieces of inspection result information 11a and the N pieces of processing history information 11b shown in FIG. The ranking result by (M × N) statistics (for example, variance ratio F value) indicating the degree of influence between the two elements is displayed by regression analysis. That is, the parameter, F value, number of groups, and number of data are displayed with the inspection result information as the objective variable and the processing history information as the explanatory variable.
The graph display unit 32 displays a graph representing the relationship between the two elements, for example, a boxplot at the top, and simultaneously displays a trend chart at the bottom. As shown in FIG. 7B, the details of the boxplot mean the maximum value, the third quartile, the median, the first quartile, and the minimum value, respectively.
The analysis operation unit 33 includes an analysis execution unit 33a that executes an optimal production route analysis, a processing step designation unit 33b that designates the ranked processing steps, and a switching designation unit 33c that designates all models and designates a single model. Place.
The two graphs shown in the graph display unit 32 in FIG. 7 represent the relationship between the performance 1 that is the inspection result information selected by the ranking result display unit 31 and the process 7 that is the processing history information. A statistic (dispersion ratio F value) representing the degree of influence between elements means 2.35.

The user designates the analysis execution unit 33a on the screen of the initial state of the optimum production route analysis template shown in FIG.
“Specify” on the optimum production path analysis template in FIG. 8A is a mode in which, for example, the cursor is moved to a certain cell or switch with the mouse and is specified by clicking. By this operation, multiple regression analysis is performed using the first M pieces of examination result information 11a as objective variables and N pieces of processing history information 11b as explanatory variables. As a result of the multiple regression analysis, ranking is performed based on (M × N) statistics (for example, variance ratio F value) indicating the degree of influence between two elements. The ranking result is displayed on the ranking result display unit 31. That is, as shown in FIG. 8B, the inspection result information is “performance 6”, the processing history information is “process name 1”, the parameter is “process name 1”, the F value is “14.25”, and the number of groups is “ 3 ”and the number of data“ 120 ”are displayed.
The graph display unit 32 shown in FIG. 8B shows the relationship between the performance 6 that is the inspection information that needs the most attention and the process name 1 that is the processing history information that most affects the performance 6, and the box plot and trend. Represented as a chart.

The inspection result information and the processing history information in the ranking result display unit 31 shown in FIG. 8B are ranked by analyzing the relationship (M × N), and the result is shown in FIG. 8C. It can be confirmed from the pull-down list. At the same time, when the user designates arbitrary inspection result information and processing history information, the relationship between the inspection result information and the processing history information can be displayed on the graph display unit 32 to easily draw a graph.
The box-and-whisker diagram in FIG. 8B shows that the performance 6 that is the inspection result information is affected by the devices (SEM 14, SEM 16, and SEM 17) included in the process name 1 whose variation factor is the processing history information. Represents.

In this way, it is possible to more easily extract and narrow down the production process that maximizes the number of products to be analyzed (inspection result information), and the intention of which process in which process should be prioritized. It makes it possible to facilitate the decision.
Here, the inspection result information indicating the product performance of the product includes the current consumption, processing speed, durability, etc. of the device electrical characteristics test, and the processing shape (oxide film thickness, Line width, overlay value, foreign matter inspection value, pattern defect value) can also be used.
Further, the processing information or processing history information to be used can be used not only for the manufacturing process and the manufacturing apparatus, but more specifically, a processing chamber or a processing position in a furnace such as a diffusion furnace in each processing apparatus. .
As described above, the inspection result and the processing history information can be displayed on the result display unit 31, the graph display unit 32, and the analysis execution unit 33 for each model, processing process, and manufacturing device. It can be grasped and verified in real time. In addition, the above operations are updated in a certain period, and the predicted value and actual value when using the production apparatus of each assigned production route are constantly monitored, and if deviation occurs, the variation is reduced. Correction processing is performed.
In the correction process, for example, the processing history information uses information for the past three months from the evaluation time (for example, the current day), calculates the optimum production route using the processing history information for the past three months, and makes such an optimum The production route calculation is performed so as to reduce variation by recalculating the production route calculation every day while renewing it every day. In this way, the mechanism that can update the processing history information for calculating the optimum production route for a certain period (for example, three months) and reflect the processing history up to immediately before the evaluation time is a correction process that reduces variations. It is effective as

(Second Embodiment)
In the second embodiment, the inventory increase loss or the amount of input material generated when the scheduled production schedule or the scheduled delivery quantity to which the production lot of the product is allocated, the production date on the actual production line, and the number of production is generated. This is a production line evaluation method in which the additional input loss caused by the difference between the actual production amount or the production loss that is the sum of the inventory increase loss and the additional input loss is the second analysis target. The production line of the second embodiment is the same as in FIG. 2, the production apparatus is the same as in FIG. 3, and the production line evaluation system is the same as in FIG.

生産過程での生産ロスを最小化する生産経路とは、やはり各製造工程の生産性の高い装置で生産する経路である。生産過程全体の最適化とは、顧客要求である出荷予定日程と出荷予定数量を維持しながら、生産能力の限られた製造装置から優先順位をつけて各工程の有効な装置を選択して全ての機種で生産ロスを最小化する生産経路である。 The production loss that is the second analysis target is the inventory increase loss, the additional input loss, or the production loss that is the sum of the inventory increase loss and the additional input loss. That is, as shown in Table 2, the production loss for each model is calculated by summing the production loss for each product lot obtained by adding the inventory increase loss, the additional input loss, or the inventory increase loss and the additional input loss. Furthermore, the production loss of the entire product is calculated by summing up the production loss for each model.

The production path that minimizes the production loss in the production process is also a path for production with a highly productive device in each manufacturing process. The optimization of the entire production process means that the effective schedule for each process is selected by prioritizing the manufacturing equipment with limited production capacity, while maintaining the planned shipping schedule and the planned shipping quantity, which are customer requirements. This is a production path that minimizes production loss with this model.

As shown in FIG. 5, the flowchart of the second embodiment is characterized by having a model-specific plan input step S7. Other steps are the same as those described in the first embodiment.
That is, in the model-specific plan input step S7, model-specific plan information such as the priority of each model, the planned shipping schedule, and the planned shipping quantity is input from the variable selection operation unit 13. The optimum production route analysis step S4 analyzes the optimum production route from the association information in the information association step S3 and the model-specific plan information in the model-specific plan input step S7.

FIG. 9 shows an example of the model-specific plan information data table input in the model-specific plan input step (S7).
In the second embodiment, the information used for calculating the production loss that is the second analysis target of the optimum production route extraction is edited from the information input as customer information. As shown in FIG. 9, the model-specific plan information data table includes model type information 41b and model information 41b and customer information 41c of model-specific priority, scheduled shipping date, planned shipping quantity for each lot number 41a that is the product identification of the product. Corresponding. This model-specific plan information data table is edited together with the association data table of FIG. 6 in the data editing function 12a of the optimum production route analysis template 12 shown in FIG. 4 using L identifiers (lot numbers) as keys. Used as data. That is, all the N pieces of processing history information 11b shown in FIG. 6 are used as explanatory variables, and the difference between the model-specific plan information shown in FIG. 9 and the shipping schedule originally required from the actual processing history information. The target variable is the sum of the inventory increase loss and the additional input loss that occurs.
Based on this information, the multiple regression analysis by the production line evaluation system equipped with the production line evaluation method shown in FIG. 4 makes it easier to narrow down the extraction of the production process that minimizes the production loss that is the second analysis target. This makes it possible to facilitate the decision of which device in which process should be processed with priority.

(Third embodiment)
The third embodiment is a production line evaluation method in which a case of a delivery schedule difference that is a defect from a shipping schedule to which a production lot of a product is allocated is a third analysis target. The production line of the third embodiment is the same as in FIG. 2, the production apparatus is the same as in FIG. 3, and the production line evaluation system is the same as in FIG.

  The third delivery target difference in delivery schedule is a defect from the scheduled shipping schedule to which the production lot is assigned. This index is expressed as “1” when the shipping schedule is satisfied for each production lot. For example, when the scheduled shipping schedule can be achieved, the value is smaller than “1”, and when the scheduled shipping schedule cannot be achieved, the value is larger than “1”. The delivery schedule difference for each model is also expressed as a product for each production lot, and the state satisfying the schedule is expressed as “1”. That is, as shown in Table 3, the difference in the delivery schedule for each product lot is calculated as the difference with respect to the reference value “1” to calculate the delivery schedule difference for each model. Furthermore, the difference in the delivery schedule for each product is calculated by summing up the delivery schedule differences for each model.

生産過程での納期計画差を最小化する生産経路とは、生産能力の限られた製造装置から優先順位をつけて各工程の有効な装置を選択して全ての機種で納期計画差を最小化する生産経路である。
第３実施形態のフローチャートは、第２実施形態のフローチャートと同じである。

The production route that minimizes the delivery schedule difference in the production process is to prioritize the production equipment with limited production capacity and select the effective equipment for each process to minimize the delivery schedule difference for all models. Production route.
The flowchart of the third embodiment is the same as the flowchart of the second embodiment.

FIG. 10 shows a delivery date plan difference table showing delivery date plan differences, which is information calculated as a difference between the production schedule of each production lot and the production history of the current production process and processing history information. . FIG. 10 shows the product model name, that is, the product model name 51b, the model priority, the shipping schedule, and the delivery schedule difference 51c for each lot number 51a. This delivery date plan difference table is based on the model-specific plan information data table of FIG. 9 and the linking data table of FIG. 6 by the data editing function 12a in the production line evaluation system having the production line evaluation method of FIG. Calculated. That is, all the N pieces of processing history information 11b shown in FIG. 6 are used as explanatory variables, and a delivery value initially required as a reference value “1” based on the delivery schedule difference shown in FIG. 10 and the actual processing history information. The degree of deviation from the schedule is the objective variable.
Based on this information, the multiple regression analysis by the production line evaluation system equipped with the production line evaluation method shown in FIG. 4 makes it possible to further narrow down the extraction of production processes that minimize the difference in delivery schedule, which is the third analysis target. This makes it easy to make a decision on which device in which process should be preferentially processed.

(Fourth embodiment)
In the above, the production line evaluation system provided with the production line evaluation method has been described by analyzing the production process of one model, but this fourth embodiment has a plurality of production lines and products of a plurality of models. It is explained that it is also effective when producing by mixing the two.
As an example, as shown in FIG. 11A, a case will be described in which two models A and B are produced on the model A production line and the model B production line. Here, processes using the same equipment have the same process name and are defined as the entire production line. A plurality of manufacturing apparatuses are arranged in at least one process of each production line.

That is, in the example of FIG. 11A, the production line for model A does not have step 3, and the production line for model B does not have step 2 and step 5. And process 1 and process 4 are equipped with the manufacturing equipment of No. 1 to No. 3, and since two models are produced in process 1 and process 4, it seems that three manufacturing apparatuses are effectively used with two models. I have to. Step 2 and Step 3 include Unit 1 and Unit 2, respectively. In step 5, one manufacturing apparatus is shared to manufacture models A and B.
The production capacity of each production apparatus provided in each of these processes is as shown in FIG. 11B, and the production apparatus of process 1 has a production capacity of 100 for both No. 1 to No. 3. The production apparatus of process 2 has a production capacity of 200 for both No. 1 and No. 2. The production apparatus of the process 3 has a production capacity of 300 for both No. 1 and No. 2. The production apparatus of process 4 has a production capacity of 100 for both No. 1 to No. 3. The production apparatus of the process 5 has a production capacity of No. 1 machine of 500.

In the production process, both the model A and the model B are produced in the process 4, and the apparatus 1, the apparatus 2, and the apparatus 3 are arranged in the process. The relationship between the product performance Q and the machining shape process parameter X in step 4 has a strong correlation as shown in FIG. Next, when classified by device, as shown in FIG. 1 (b), it can be seen that there are many defects in which the product performance Q deviates from the product performance management standard R in the products processed by the second machine. Furthermore, when the same processing conditions are used in the same process, the model A and model B, as shown in Fig. 1 (c), have problems only with model A even when processed by the same second machine. I understand. As a result, when the product performance Q is to be controlled within the product performance management standard R, it can be seen that it is most effective to take measures against the second machine that is the cause of failure in the model A. If the total production volume of process 4 for model A and model B is less than the total production capacity of the three manufacturing devices, and if the production of model A can be distributed to devices other than unit 2, unit 2 It should be given top priority not to carry out production.
The production line evaluation system having the production line evaluation method of the present invention has a process that affects inspection result information (for example, the number of shipments, production loss, delivery schedule difference) that is the object of analysis due to the difference in the devices of each process. It can be easily extracted, and it can be easily simulated which apparatus is used for production.

The relationship figure of the product performance of a semiconductor device and process data is shown. The process drawing of the production line used as the object of the present invention is shown. The whole block diagram of a production apparatus provided with the production line evaluation system of this invention is shown. The block diagram of the production line evaluation system provided with the production line evaluation method of this invention is shown. The flowchart figure of the production line evaluation method including the optimal production path analysis of this invention is shown. The figure with a string data table used by this invention is shown. An explanatory view of a layout example of an optimal production path analysis template used in the present invention is shown. The template figure after the optimal production path calculation from the initial state of the optimal production path analysis template used by this invention, the explanatory drawing of a box whip figure, and the detailed explanatory drawing of the pull-down list of the ranking result display part are shown. The model-specific plan information data table figure used in 2nd Embodiment of this invention is shown. The delivery date plan difference table figure used in 3rd Embodiment of this invention is shown. The whole block diagram of the production line explaining the 4th embodiment of the present invention is shown.

Explanation of symbols

1 Production Equipment 2 CIM System 3 Customer Information Management System 4 Production Line Evaluation System 5 LAN
11 Linking data table 12 Optimal production route analysis template 12a Data editing unit 12b Multiple regression analysis function and ranking unit 12c Processing step designation unit 12d Ranking repetition unit 12e Graph display unit 13 Variable selection operation unit 14 Production route analysis operation unit 15 Display unit 30 Optimal production route analysis template 31 Ranking result display unit 32 Graph display unit 33 Analysis operation unit

Claims (18)

In a production line having a plurality of production paths, an evaluation method for evaluating the productivity of the plurality of production paths,
A processing information selection step for selecting from a processing information storage unit that stores processing information when a product is produced on the production route;
An inspection information collecting step for collecting from an inspection information storage unit for storing inspection information of products produced in the production path;
Linking the collected processing information and the inspection information to obtain link information; and
A ranking step of ranking the plurality of production paths in order of productivity based on the association information;
A production route display step for displaying the ranked production routes on a display device;
A production line evaluation method comprising:   The production line evaluation method according to claim 1, wherein the association step associates the processing information and the inspection information for each production lot of the product.   The ranking step includes a production route analysis step for analyzing the plurality of production routes based on the association information, and an analysis result ranking step for ranking the plurality of production routes based on the analyzed result. The production line evaluation method according to claim 1 or 2, comprising:   4. The production line evaluation method according to claim 3, wherein the plurality of production paths are respectively ranked for a plurality of product models by repeating the production path analysis step, the analysis result ranking step, and the production path display step. .   The product further includes an input unit for product plan information including a priority for each model of the product, a planned shipping schedule or a planned shipping quantity, and the ranking step includes a plurality of productions based on the linking information and the product plan information. The production line evaluation method according to claim 1, wherein the routes are ranked.   The production line evaluation method according to claim 1, wherein the ranked production route is a production route that maximizes the number of shipments calculated based on the production capacity of each production route.   The production line evaluation method according to claim 1, wherein the ranked production path is a production path that maximizes the number of shipments for each model and maximizes the total number of shipments of all models.   The production line evaluation method according to claim 1, wherein the ranked production path is a production path that minimizes a production loss for each model and minimizes a production loss for all models.   2. The production line evaluation method according to claim 1, wherein the ranked production paths are production paths that minimize a priority, a scheduled delivery date or a planned shipment quantity for each model, and a production loss for each production process.   The production line evaluation method according to claim 1, wherein the ranked production paths are production paths that minimize a delivery schedule difference for each model and minimize a delivery schedule difference for all models.   The production according to claim 1, wherein the ranked production route is a production route that minimizes a delivery schedule difference based on a priority for each model, a scheduled delivery schedule, a planned shipment quantity, and a production capacity for each production process. Line evaluation method.   2. The production line according to claim 1, wherein the ranked production paths are production paths in which the processing information and the inspection information are extracted in order from important items having a large influence based on simulation information obtained by a multiple regression analysis method. Evaluation methods.   The ranked production route is a production route that is updated in a certain period and has a correction process that constantly monitors the predicted value and actual value when using the manufacturing apparatus of each assigned process, and reduces variations. The production line evaluation method according to claim 1.   2. The production according to claim 1, wherein the display means displays the predicted value and actual value for each production route for every model, model, and production route so that the production control status can be grasped and verified in real time. Line evaluation method.   14. A production line control method further comprising a CIM for managing the production line, wherein the production path is controlled by the CIM according to the production path ranked by the production line evaluation method according to any one of claims 1 to 13. .   The production line evaluation program for making a computer perform each step of any one of Claim 1 to 4.   The computer-readable recording medium which recorded the production line evaluation program of Claim 16. Multiple production paths in the production line;
A processing information selection unit that selects from a processing information storage unit that stores processing information when a product is produced through the production path;
An inspection information collection unit that collects from an inspection information storage unit that stores inspection information of products produced in the production path;
A linking unit that links the collected processing information and the inspection information to obtain linking information;
A ranking unit that analyzes and ranks the plurality of production paths based on the association information;
A production line evaluation system comprising: a production route display unit for displaying the ranked production routes on a display device.

Images