JP2008192745A - Yield monitoring system and yield monitoring method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a yield monitoring system that can be utilized for developing a semiconductor device of high yield. <P>SOLUTION: A data collecting part 10 collects implementation conditions (design condition, manufacturing condition, test condition and the like), and implementation results (circuit and layout information obtained by design, test results (defective article occurrence rate) and the like), in a design process, and test process as well as a manufacturing process of a semiconductor device. An analyzing part 11, based on the implementation conditions and implementation results that have been collected, analyzes the implementation conditions or combination of them which affects a yield, and outputs the analyzing result. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a yield monitoring system and a yield monitoring method for specifying a factor that deteriorates the yield of a semiconductor device.

  In the conventional semiconductor device manufacturing process, in order to improve the yield, the manufacturing department has been trying to suppress the deterioration of the yield mainly due to adhesion of dust or the like during wafer fabrication. On the other hand, the design departments and the like faced downsizing, shortening TAT (Turn Around Time), and reducing costs, and their awareness of improving yields was weak.

  However, the manufacturing technology cannot keep pace with the recent miniaturization of semiconductor devices, and it is necessary to take measures to improve the yield even outside the manufacturing department such as the design department and the test department.

Conventionally, a technique for estimating a device that causes a deterioration in yield (for example, refer to Patent Document 1) and a technique using data mining to analyze the yield (for example, Patent Document 2,
3).
JP 2005-277290 A JP 2002-324206 A Special table 2006-506800 gazette

  However, in the development of semiconductor devices, there are a large number of processes, and for each process, there are several hundred implementation conditions (design conditions, manufacturing conditions, test conditions, etc.). And since a semiconductor device is manufactured with such a combination of implementation conditions, it is very difficult to specify where there is a factor that deteriorates the yield.

  The present invention has been made in view of these points, and an object of the present invention is to provide a yield monitoring system and a yield monitoring method that can be used for developing a semiconductor device with a high yield.

  The inventors of the present invention have a data collection unit that collects execution conditions or execution results in each of the design, manufacture, and test processes of a semiconductor device, and affects the yield based on the collected execution conditions or execution results. Providing a yield monitoring system comprising: an analysis unit that analyzes the given execution conditions or a combination of the execution conditions and outputs an analysis result.

  According to the above configuration, based on the implementation conditions or implementation results in the design, manufacture, and test processes of the semiconductor device collected by the data collection unit, the analysis unit performs the implementation conditions or implementation conditions that affect the yield. The combination is analyzed and the analysis result is output.

  The present invention collects not only the manufacturing process of semiconductor devices but also the implementation conditions and implementation results in the design process and test process, and analyzes the implementation conditions or combinations of implementation conditions that affect the yield based on the collected information. Therefore, it is possible to clarify the cause of the deterioration of the yield.

  Then, by changing the clarified implementation condition or the combination that causes the yield deterioration, it becomes possible to develop a semiconductor device with a high yield.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a schematic configuration diagram of a yield monitoring system according to the present embodiment.
The yield monitoring system includes a data collection unit 10, an analysis unit 11, and a deficient data generation unit 12.

  The data collection unit 10 collects implementation conditions and results in each of the semiconductor device design processes 1 to i, the manufacturing processes 1 to j, and the test processes 1 to k. For example, the design processes 1 to i include a logic circuit design process and a layout design process. The implementation conditions include a layout tool to be used, a type of macro cell, a chip size, an OPC (Optical Proximity Correction) correction condition, an exposure condition, a manufacturing apparatus to be used, and the like. The implementation results include, for example, circuits obtained by design, layout information, and test results (defect product occurrence rate, etc.).

  In addition, the data collection unit 10 collects the execution conditions used in each step and the obtained execution results every time the design steps 1 to i, the manufacturing steps 1 to j, and the test steps 1 to k are finished. Store in a database (not shown).

  The analysis unit 11 analyzes the information collected by the data collection unit 10. Specifically, the execution conditions and the combination of the execution conditions that affect the yield are analyzed, and the analysis result is output. The analysis is performed by data mining and defective product generation factor analysis processing.

  Data mining is a technique for finding patterns and relationships between items from a large amount of collected information. In the yield monitoring system according to the present embodiment, the relationship between the yield and the execution condition or a combination of the execution conditions is examined and output based on the collected execution conditions and execution results. For example, when the manufacturing apparatus A and the manufacturing apparatus B are used as a combination of implementation conditions, the relevance such that the defective product occurrence rate is increased or conversely decreased is examined.

  The defective product generation factor analysis process extracts the implementation conditions and the combinations thereof that are frequently used in the product group with a high defective product generation rate and the low product group, and analyzes the defective product generation factor.

  If the data collection unit 10 cannot collect the execution conditions and results necessary for the yield analysis when the analysis unit 11 performs the analysis, the insufficient data generation unit 12 obtains the necessary data by simulation or data analysis. The data is generated and supplied to the analysis unit 11 via the data collection unit 10.

  According to such a yield monitoring system, not only the manufacturing conditions in the manufacturing process, but also the execution conditions and execution results in the design process and the test process are collected, and the execution conditions and the combinations that affect the yield are analyzed. Yield deterioration factors become clear.

Hereinafter, the yield monitoring system of this embodiment will be described more specifically.
FIG. 2 is a specific system configuration example of the yield monitoring system.
The yield monitoring system includes a data collection computer 20, a database 20a, a data mining computer 21, a defective product generation factor analysis computer 22, process management computers 23-1, 23-2,..., 23-n, and a simulation computer. 24. A data processing computer 25 is provided.

Each computer and the database 20a are connected by a network such as a LAN (Local Area Network) or a WAN (Wide Area Network).
Here, the data collection computer 20 corresponds to the data collection unit 10 of FIG. 1, and the data mining computer 21 and the defective product occurrence factor analysis computer 22 are the analysis unit 11, the simulation computer 24, and the data processing computer. Reference numeral 25 corresponds to the insufficient data generation unit 12. More specifically, each function of FIG. 1 is realized by executing a program stored in a storage unit of each computer under the control of a CPU (Central Processing Unit) of each computer.

  The process management computers 23-1 to 23-n collect data on the execution conditions and results in each process such as the assembly process and the shipping process as well as the design process, manufacturing process, test process, and the like shown in FIG. To the computer 20. One process management computer may manage a plurality of processes.

The database 20a stores data collected by the data collection computer 20.
Note that the simulation computer 24 and the data processing computer 25 may be a single computer or a plurality of computers. The same applies to the other computers in FIG. For example, the data mining computer 21 and the defective product occurrence factor analysis computer 22 may be combined into a single yield analysis computer.

The operation of the yield monitoring system in FIG. 2 will be described below.
FIG. 3 is a flowchart showing a process flow of the yield monitoring method.
First, the data collection computer 20 collects and stores the execution conditions and execution results transferred each time the process is completed by the process management computers 23-1 to 23-n (step S1). For example, when the logic circuit design process is completed, before proceeding to the next layout design process, the execution conditions such as the type of the selected device and the logic circuit design tool used are selected from the process management computer that manages the logic circuit design process. In addition, the designed logic circuit information, the total number of used devices, the cost and other implementation results are transferred to the data collection computer 20. In the test process, the defective product occurrence rate, cost, and the like are transferred from the process management computer that manages the test process to the data collection computer 20 as an execution result. Further, an assembly time, a transportation cost, a transportation route, a transportation time, and the like are transferred from the process management computer that manages the assembly process and the shipping process to the data collection computer 20. The data collection computer 20 stores the collected information in the database 20a.

  Next, the CPU (not shown) of the data collection computer 20 compares the list of items that can be collected defined for each process with the actually collected information to analyze the implementation conditions that affect the yield. It is determined whether or not there is a lack of information necessary for the operation (step S2). For example, in the mask layout process, the occupation rate information occupied by the dummy pattern may not be collected when normal processing is performed. However, since this information is an element that affects the yield, it is added to the list of items that can be collected.

FIG. 4 is an example of a list of items that can be collected.
Information item name, value range, value unit, etc. are described for each process.
Such a list is stored, for example, in storage means such as a RAM (Random Access Memory) or HDD (Hard Disk Drive) (not shown) of the data collection computer 20, and is read out as necessary.

  If it is determined in step S2 that there is missing information, the data collection computer 20 requests the simulation computer 24 or the data processing computer 25 to generate the missing information. . Whether to use the simulation computer 24 or the data processing computer 25 depends on the type of information requested. Moreover, you may make it use both.

  In response to the request, the simulation computer 24 or the data processing computer 25 selects a tool (simulation software or data processing software) for acquiring the missing information (step S3). For example, when the above-described dummy pattern occupancy rate information is insufficient, a mask layout tool is selected.

  Thereafter, the tool is activated to acquire the shortage information (step S4). At this time, the simulation computer 24 or the data processing computer 25 obtains information necessary for generating insufficient information from the data collection computer 20 for the selected tool. For example, when acquiring the occupancy rate information of the dummy pattern, other execution conditions and execution results collected by the data collection computer 20 regarding the mask layout process are input to perform simulation and data processing. Information generated by simulation or data processing is acquired by the data collection computer 20.

  Note that the simulation computer 24 and the data processing computer 25 have a function of determining whether or not each tool has a newly obtainable item when selecting a tool according to the lack information requested to be obtained. Yes. This is because when the number of items that can be acquired increases due to the version upgrade of the tool, it is learned and used in the next automatic tool selection.

  Next, the data collection computer 20 determines whether or not the acquisition of the shortage information is completed (step S5). If the acquisition is not completed, the data collection computer 20 requests the simulation computer 24 or the data processing computer 25 again. Then, another tool is selected (step S6). Thereafter, the process returns to the process of step S4, and the simulation computer 24 or the data processing computer 25 activates the selected tool and similarly performs a process of acquiring insufficient information.

  If it is determined in step S5 that the data collection computer 20 has completed the acquisition of the deficient information, the data collection computer 20 additionally stores the acquired information in the database 20a (step S7), and step S2 Return to the process.

  In the process of step S2, if the data collection computer 20 determines that there is no missing information, the data mining computer 21 and the defective product occurrence factor analysis computer 22 store the information stored in the database 20a. In addition, data analysis such as data mining and extraction of defective product occurrence factors is performed (step S8), and the result is output (step S9).

Details of the data analysis process will be described below.
The data mining computer 21 performs data mining on the basis of information (implementation conditions and implementation results) stored in the database 20a, and the relationship between products with low defective product occurrence rates and implementation conditions, and conversely defective products. Analyze the relationship between products with high incidence and implementation conditions, or the relationship between cost and TAT and implementation conditions. For example, a product using the exposure apparatus B with a chip size of A has a low defective product generation rate, and a product using the layout tool C has a long TAT until the reticle shipment, and other processes that are performed simultaneously. Similarly, the product is examined for a longer TAT and higher costs.

  The analysis result is provided to the user by displaying it on a display (not shown) of the data mining computer 21 or transmitting it to another computer connected via a network as necessary by e-mail or the like.

Next, a defective product generation factor analysis process by the defective product generation factor analysis computer 22 will be described.
FIG. 5 is a flowchart showing the flow of defective product occurrence factor analysis processing.

  First, the defective product generation factor analysis computer 22 determines whether or not the defective product occurrence rate of a product is a specified value (for example, 1%) or more (step S10). Here, if it is lower than the threshold value, the defective product occurrence factor analysis process is terminated. As a result, it is possible to avoid performing an analysis process that consumes time and CPU resources of a computer even for a product having a low defective product occurrence rate that does not require analysis.

  If it is determined in step S10 that the defective product occurrence rate of the product is equal to or higher than the threshold value, the product group having the defective product occurrence rate equal to or higher than a specified value (for example, 10%) is frequently used in common. The group g1 composed of the execution conditions and the combination thereof is extracted from the database 20a and output to a display (not shown) of the defective product occurrence factor analysis computer 22 (step S11).

  Next, a group g2 consisting of implementation conditions and combinations thereof commonly used in product groups with a defective product incidence smaller than the prescribed value is extracted from the database 20a and output in the same manner as in step S11 (step S11). S12).

  Thereafter, the implementation conditions of the group g1 extracted in the process of step S12 are excluded from the implementation conditions of the group g1 extracted in the process of step S11, and the remaining implementation conditions are changed in the semiconductor device to be developed. The group g3 consisting of implementation conditions that do not cause a serious problem (change is permitted) is extracted as a candidate for a defective product generation factor and output (step S13). This narrows down the cause of defective products.

  Next, the defective product generation factor analysis computer 22 performs a simulation of how the defective product generation rate changes by appropriately changing the execution conditions of the extracted group g3 (step S14).

  Then, the group g4 composed of the implementation condition that minimizes the defective product occurrence rate and the combination thereof is extracted and output (step S15). Here, the change (transition) of the defective product occurrence rate when the implementation condition is changed may be output as a graph.

Further, the execution condition of the group g4 is notified to the process management computers 23-1 to 23-n of each process by e-mail or the like (step S16).
Thereby, in each process, when manufacturing the same product next time, it can process using the implementation conditions that the defective product rate occurrence rate becomes low, and the yield of the semiconductor device can be improved. .

  It should be noted that for all the implementation conditions of group g1 and group g2 and the combinations thereof, the upper limit (or lower limit) of the implementation condition value of group g1 having a high defective product occurrence rate is handled by group g2 having a low defective product occurrence rate. The value of the execution condition may be set as a lower limit (or upper limit), and the numerical value may be varied within the range, and a change in the defective product occurrence rate at that time may be analyzed and output.

For example, if the chip area is 200 mm ² as the implementation condition (manufacturing condition) of the group g1 and 230 mm ² in the group g2, the chip area is changed in a range of 200 to 230 (actually, the chip with the same function). Adjust by changing to a macro cell with a different area). As a result, the optimum chip area (macro cell) can be determined, and information useful for improving the defective product occurrence rate in the future can be obtained.

  Further, the defective product generation factor analysis computer 22 compares the past defective product generation rate of the same product with the current defective product generation rate, and analyzes the defective product generation factor from the execution conditions used in each. You may do it.

FIG. 6 is a flowchart showing a flow of defective product occurrence factor analysis processing by comparison with the same product in the past.
First, the defective product generation factor analysis computer 22 searches the database 20a to determine whether the same product as the product produced this time has been produced in the past, such as six months ago or one year ago. (Step S20). If there is no product made in the past, the process is terminated, and if there is the product, the defective product generation rate when it was manufactured in the past and the defective product generation rate when manufactured this time It is determined whether there is a difference (step S21). At this time, the defective product occurrence factor analysis computer 22 determines that there is a difference when the defective product occurrence rate difference is equal to or greater than a predetermined threshold (for example, 0.01%). Thereby, consumption of time and CPU resources of the computer can be suppressed. If it is determined that there is no difference, the processing is terminated as it is.

  If it is determined that there is a difference in the rate of defective products, whether or not the implementation conditions used for the product manufactured this time differ from the implementation conditions used for the product manufactured in the past Is determined by searching the database 20a, and it is further determined whether or not the current product has an improved defective product generation rate compared to past products (step S22). Here, if the implementation conditions are the same for the past and the current one, or if the defective product occurrence rate is not improved, the processing is terminated as it is.

  The search result and the determination result may be output and displayed on the display of the defective product occurrence factor analysis computer 22. Thereby, in the process of this product, it can be clarified which defective product occurrence rate has improved by changing which execution condition.

If there is a difference in the execution conditions in the process of step S22 and the defective product occurrence rate of the current product is lower than that of the past product, the process proceeds to step S23.
In the process of step S23, among the execution conditions used in the manufacture of the same product in the past, the execution conditions different from the execution conditions used in the manufacture of the current product found in the process of step S22 (hereinafter referred to as implementation) A similar product using the condition A) is searched from the database 20a (step S23). However, similar products that are permitted to be changed in the execution condition A are searched. There may be a plurality of implementation conditions A. The similar products are, for example, products of the same series or products having different implementation conditions at only a few locations.

  Thereafter, the defective product occurrence factor analysis computer 22 changes the implementation condition A of the similar product to the implementation condition of the current product that has been an improvement factor of the defective product occurrence rate and changes the defective product occurrence rate. Is simulated (step S24). The simulation result here may be output and displayed on the display of the defective product occurrence factor analysis computer 22.

  As a result of the simulation, when the defective product occurrence rate is improved by changing the execution condition A, the changed execution condition and the combination thereof are used as a process management computer 23-1 for each process in manufacturing a similar product. To 23-n by e-mail or the like (step S25).

Thereby, in each process, it can avoid using the implementation conditions and the combination which deteriorate the defective product rate generation rate, and can improve the yield of a semiconductor device.
By the way, by using the yield monitoring system of the present embodiment, it is possible to analyze the implementation conditions and their combinations that affect the cost.

  The cost is calculated for each process, transferred from the process management computers 23-1 to 23-n to the data collection computer 20, and collected. Based on the collected cost information, analyze the cost deterioration factors in each process. The cost analysis may be performed by, for example, the data mining computer 21 or the defective product occurrence factor analysis computer 22, but in the following, a cost analysis dedicated computer (not shown) that can refer to the database 20a is used. Will be described. Note that this computer also belongs to the analysis unit 11 in FIG. 1 that analyzes collected information.

FIG. 7 is a flowchart showing a flow of cost increase factor extraction processing in each step.
In a certain process A, the cost analysis dedicated computer searches the database 20a for a group g5 of products having a higher cost than the sales forecast (step S30). And the group g6 which consists of the implementation conditions with a high frequency currently used in common with the group g5, and its combination is extracted (step S31).

  Next, in the process A, the cost analysis dedicated computer searches the database 20a for a group g7 of products whose cost is lower than the sales forecast (step S32). And the group g8 which consists of the implementation conditions with high frequency currently used in common by the group g7, and its combination is extracted (step S33).

  Thereafter, the implementation conditions of the group g6 extracted in the process of step S33 are excluded from the implementation conditions of the group g6 extracted in the process of step S31, and the remaining implementation conditions are changed when the product to be manufactured is manufactured. Even if it does not cause a serious problem (change is permitted), a group g9 of implementation conditions is extracted and output as a candidate for a cost increase factor (step S34). As a result, the cause of the cost increase is narrowed down.

  Then, the cost of the group g9 is changed as appropriate to simulate a change in cost (step S35). From the simulation result, the group g10 consisting of the implementation condition and the combination thereof with the lowest cost is extracted and output (step S36). In addition, you may make it output the graph of the change of the cost according to the change of implementation conditions.

Finally, the execution condition of the extracted group g10 is notified to the process management computer of process A by e-mail or the like (step S37).
For all the implementation conditions of group g6 and group g8 and combinations thereof, the upper limit (or lower limit) of the implementation condition value of group g6 with high cost and the corresponding implementation condition value of group g8 with low cost are set. As the lower limit (or upper limit), the numerical value may be varied within the range, and the transition of the cost at that time, the combination of the values with the lowest cost, or the like may be output, or the process A may be notified.

By performing the above process on each process, the process of each process can be performed using the implementation conditions and the combination effective for cost reduction.
By the way, since the cause of the cost increase is analyzed for each process as described above, the actual cost may be higher than expected. For this reason, the factor of cost increase may be extracted again based on the total cost in the final process (for example, after the completion of the shipping process).

FIG. 8 is a flowchart showing the flow of the process for extracting the cost increase factor in the final process.
The cost analysis dedicated computer first determines whether or not the process of the target product is the final process (step S40), and does not perform the following processing until the final process is reached.

  When it is determined as the final process, the cost analysis dedicated computer extracts the total cost of the target product from the database 20a, and determines whether the total cost is higher than the sales forecast (step S41). If it is not expensive, the process is terminated as it is.

  If it is determined that the cost is high, a similar product (for example, a product of the same series, a product with a small difference in implementation conditions) having a lower cost (total cost) than the target product is searched from the database 20a (step S42). ).

  Subsequently, the cost analysis dedicated computer determines whether or not the execution condition used in the similar product having a lower total cost than the target product can be used in the target product (step S43). Here, it is determined whether or not the execution condition in the production of the target product may be changed to the same as that of a similar product with a low total cost. If it is not usable, the process is terminated.

If there is a usable execution condition, the execution condition of the target product is appropriately changed to a similar product with a low total cost, and a simulation is performed (step S44).
As a result of the simulation, the execution conditions with the lowest total cost and combinations thereof are output to, for example, the display of a computer dedicated to cost analysis, and further notified to the process management computers 23-1 to 23-n of each process by e-mail etc. (Step S45).

Thereby, the combination of the implementation conditions for manufacturing the same product next at low cost can be clarified.
Of the target products and similar products with lower cost, the upper limit (or lower limit) is the implementation condition value of the target product, and the lower limit (or upper limit) is the corresponding implementation condition value for similar products with lower costs. The numerical value may be varied within the range, and the transition of the cost at that time, the combination of the values with the lowest cost, or the like may be output or each process may be notified.

  Note that the above-described data mining, defective product generation factor analysis processing, and cost increase factor extraction processing may be used in combination. By combining them, it becomes possible to analyze in more detail and from various fields.

(Supplementary Note 1) A data collection unit that collects implementation conditions or implementation results in each of the steps of designing, manufacturing, and testing a semiconductor device;
Based on the collected execution conditions or the execution results, an analysis unit that analyzes the execution conditions or a combination of the execution conditions that affect the yield and outputs the analysis results;
A yield monitoring system characterized by comprising:

  (Supplementary Note 2) The yield according to Supplementary Note 1, wherein the analysis unit performs data mining on the implementation condition or the implementation result, and analyzes and outputs the relationship between the yield and the implementation condition or the combination. Monitoring system.

  (Additional remark 3) The said analysis part is the product group smaller than the said 1st group which consists of the said implementation conditions or the said combination currently used in common with the product group whose defective product incidence is more than a predetermined regulation value, and The yield monitoring system according to appendix 1 or 2, wherein a factor causing defective products is specified by comparing with the second group consisting of the implementation condition or the combination commonly used in the above.

  (Supplementary Note 4) The analysis unit determines a third group consisting of the implementation conditions that are allowed to be changed among the implementation conditions excluding the second group from the first group as a cause of defective products. The yield monitoring system according to appendix 3, wherein the yield monitoring system outputs the candidate.

  (Additional remark 5) The said analysis part changes the said implementation conditions of the said 3rd group, simulates for every product, and analyzes the change of the said defective product incidence rate, The yield monitoring of Additional remark 4 characterized by the above-mentioned system.

(Supplementary Note 6) The yield monitoring system according to Supplementary Note 5, wherein the analysis unit outputs the implementation condition or the combination that minimizes the defective product occurrence rate.
(Additional remark 7) The said analysis part analyzes the said defective-product generation factor, only when there exists the said product group whose said defective product incidence is more than a predetermined threshold value, Any one of Additional remark 3 thru | or 6 characterized by the above-mentioned. Yield monitoring system according to item.

  (Supplementary Note 8) The analysis unit includes a first value of a first implementation condition belonging to the first group and a second implementation condition belonging to the second group and corresponding to the first implementation condition. Any one of the second values is an upper limit and the other is the lower limit, and the simulation is performed to analyze and output the change in the defective product occurrence rate. Yield monitoring system described in

  (Supplementary Note 9) When there is a difference in the defective product occurrence rate between the product manufactured in the past and the product manufactured this time, the analysis unit determines the cause of the difference based on the implementation condition or the combination of both. The yield monitoring system according to any one of appendices 3 to 8, wherein the yield is analyzed and output.

  (Additional remark 10) The said analysis part outputs the said implementation conditions or the said combination after a change, when the said implementation conditions which are the factor of the said difference are changed, and the said defective product generation rate is improved by simulation. The yield monitoring system according to supplementary note 9, wherein

(Supplementary note 11) The yield monitoring system according to supplementary note 9 or 10, wherein the analysis is performed only when the difference is equal to or greater than a predetermined threshold value.
(Supplementary Note 12) The yield monitoring system according to any one of Supplementary notes 1 to 11, wherein the data collection unit further collects implementation conditions or implementation results in an assembly process and a shipment process of the semiconductor device.

  (Additional remark 13) The said analysis part is a cheap product compared with the 4th group which consists of the said implementation conditions or the said combination currently used in the product group whose cost is high compared with a sales forecast, and the said sales forecast. According to any one of appendices 1 to 12, characterized in that a factor of cost increase is specified by comparing with the fifth group consisting of the implementation condition or the combination commonly used in a group. Yield monitoring system.

  (Additional remark 14) The said analysis part is a factor of a cost increase to the 6th group which consists of the said implementation conditions in which the change is permitted among the said implementation conditions except the said 5th group from the said 4th group. 14. The yield monitoring system according to appendix 13, wherein the yield monitoring system is output as a candidate.

  (Additional remark 15) The said analysis part changes the said implementation conditions of the said 6th group in which change is permitted, performs simulation for every product, and analyzes the said implementation conditions or the said combination that the said cost becomes the minimum. The yield monitoring system according to appendix 14, wherein

  (Additional remark 16) The said analysis part performs the specific process of the factor of the said cost increase for every process, Then, it performs again by comparing a total cost and a sales forecast in a final process, The additional remark 13 thru | or 13 characterized by the above-mentioned The yield monitoring system according to any one of 15.

  (Supplementary Note 17) When the execution condition or the execution result necessary for yield analysis cannot be collected by the data collection unit, an insufficient data generation unit that generates the execution condition or the execution result that could not be collected The yield monitoring system according to any one of appendices 1 to 16, further comprising:

(Additional remark 18) The step which a data collection means collects the implementation conditions or the implementation result in each process of a design, manufacture, and test process of a semiconductor device,
Analyzing the execution conditions or a combination of the execution conditions affecting the yield based on the execution conditions or the execution results collected by the analysis means, and outputting the analysis results;
A yield monitoring method characterized by comprising:

It is a schematic block diagram of the yield monitoring system of this Embodiment. It is a specific system configuration example of a yield monitoring system. It is a flowchart which shows the flow of a process of the yield monitoring method. It is an example of a list of items that can be collected. It is a flowchart which shows the flow of a defective product generation factor analysis process. It is a flowchart which shows the flow of the defect generation factor analysis process by comparison with the past same product. It is a flowchart which shows the flow of the extraction process of the cost increase factor in each process. It is a flowchart which shows the flow of the extraction process of the cost increase factor at the time of the last process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Data collection part 11 Analysis part 12 Insufficient data generation part 20 Data collection computer 20a Database 21 Data mining computer 22 Defective product generation factor analysis computer 23-1 to 23-n Process management computer 24 Simulation computer 25 Data processing Computer

Claims (5)

A data collection unit for collecting implementation conditions or implementation results in each process of designing, manufacturing and testing a semiconductor device;
Based on the collected execution conditions or the execution results, an analysis unit that analyzes the execution conditions or a combination of the execution conditions that affect the yield and outputs the analysis results;
A yield monitoring system characterized by comprising:   The analysis unit is commonly used by the first group consisting of the implementation condition or the combination that is commonly used by product groups having a defective product occurrence rate equal to or higher than a predetermined specified value, and by a product group that is smaller than the specified value. 2. The yield monitoring system according to claim 1, wherein a factor causing defective products is specified by comparing with the second group consisting of the implementation condition or the combination.   The analysis unit outputs a third group consisting of the implementation conditions that are allowed to be changed among the implementation conditions excluding the second group from the first group as candidates for defective product occurrence factors. The yield monitoring system according to claim 2, wherein:   When there is a difference in the defective product occurrence rate between a product manufactured in the past and a product manufactured this time, the analysis unit analyzes and outputs the factor of the difference based on the execution condition or the combination of both. 4. The yield monitoring system according to claim 2 or 3, wherein: Data collecting means for collecting execution conditions or results in each step of designing, manufacturing and testing a semiconductor device;
Analyzing the execution conditions or a combination of the execution conditions affecting the yield based on the execution conditions or the execution results collected by the analysis means, and outputting the analysis results;
A yield monitoring method characterized by comprising:

Images