Semiconductor device characteristic value estimation system

Abstract

To provide a semiconductor element characteristic value estimation system.SOLUTION: A semiconductor element characteristic value estimation system includes an input unit, a database, and a processing unit. A first step list, a second step list, and a characteristic value of a semiconductor element are input to the input unit. The database has a function of storing a group of step lists and a group of characteristic values of semiconductor elements. The processing unit has a function of performing comparison between two step lists selected from the first step list and the group of step lists; a function of performing a test using two or more characteristic values of semiconductor elements selected from the characteristic value of the semiconductor element and the group of characteristic values of the semiconductor elements; a function of performing regression analysis of parameters for a step and two or more characteristic values of semiconductor elements selected from the characteristic value of the semiconductor element and the group of characteristic values of the semiconductor elements; and a function of estimating a characteristic value of a semiconductor element from the second step list.SELECTED DRAWING: Figure 11

Description

[Technical Field]

[0001] 1. Field of the Invention One aspect of the present invention relates to a method for estimating a characteristic value of a semiconductor device. Another aspect of the present invention relates to a system for estimating a characteristic value of a semiconductor device.

[0002] In this specification and elsewhere, a semiconductor element refers to an element that can function by utilizing semiconductor characteristics. Examples include semiconductor elements such as transistors, diodes, light-emitting elements, and light-receiving elements. Another example of a semiconductor element is a passive element such as a capacitor, resistor, or inductor that is formed by a conductive film or an insulating film. Another example of a semiconductor element is a semiconductor device that includes a circuit having a semiconductor element or a passive element. [Background technology]

[0003] In recent years, in fields using artificial intelligence (AI), robotics, and energy fields that handle high power such as power ICs, the development of new semiconductor elements has been progressing to solve issues such as increasing computational loads and power consumption. While the integrated circuits and semiconductor elements used in integrated circuits required by the market are becoming increasingly complex, there is a demand for the early launch of integrated circuits with new functions. However, the process design, device design, and circuit design in the development of semiconductor elements require the knowledge, know-how, and experience of skilled engineers.

[0004] In recent years, methods for optimizing the manufacturing process of semiconductor devices, methods for estimating device characteristics, etc. Patent Document 1 discloses a method for calculating image feature amounts from an SEM image of a cross-sectional shape pattern of a semiconductor device, and estimating the device characteristics of an evaluation pattern from the correspondence between the image feature amounts and the device characteristics. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2007-129059 Summary of the Invention [Problem to be solved by the invention]

[0006] In the manufacturing process of semiconductor devices, there are many steps required to complete a semiconductor device, and the types of steps and processing conditions vary widely. Therefore, it is difficult to verify the process dependency of characteristic values calculated from the electrical characteristics of a semiconductor device (sometimes simply referred to as the characteristic values of a semiconductor device). Furthermore, it requires a huge amount of effort to compare the process and characteristic values of a newly prototyped semiconductor device with the process and characteristic values of semiconductor devices prototyped in the past.

[0007] In the prior art, it is difficult to compare a semiconductor device manufacturing process that includes lengthy steps with the manufacturing process of a previously prototyped semiconductor device, and it is also difficult to identify which steps are unique to a newly prototyped semiconductor device.

[0008] Therefore, an object of one embodiment of the present invention is to provide a method for estimating characteristic values of a semiconductor element to be prototyped.Another object of one embodiment of the present invention is to provide a system for estimating characteristic values of a semiconductor element to be prototyped.Another object of one embodiment of the present invention is to provide a system for automatically comparing the process and characteristic values of a semiconductor element prototyped this time with the process and characteristic values of a semiconductor element prototyped in the past.

[0009] Note that the description of these problems does not preclude the existence of other problems. Note that one embodiment of the present invention does not necessarily solve all of these problems. Note that problems other than these will become apparent from the description of the specification, drawings, claims, etc., and it is possible to extract other problems from the description of the specification, drawings, claims, etc. [Means for solving the problem]

[0010] One aspect of the present invention is a semiconductor device characteristic value estimation system having an input unit, a database, and a processing unit. A first process list, a second process list, and semiconductor device characteristic values are input to the input unit. The database has a function of storing a group of process lists and a group of semiconductor device characteristic values. The processing unit has a function of comparing the first process list and two process lists selected from the group of process lists, a function of performing testing using the semiconductor device characteristic values and two or more semiconductor device characteristic values selected from the group of semiconductor device characteristic values, a function of performing regression analysis between process parameters, the semiconductor device characteristic values, and two or more semiconductor device characteristic values selected from the group of semiconductor device characteristic values, and a function of estimating the semiconductor device characteristic values from the second process list.

[0011] In the above-described system for estimating characteristic values of semiconductor devices, it is preferable to use a diff algorithm for comparison. Furthermore, it is preferable to use a t-test as a test using characteristic values of two semiconductor devices, and a non-parametric test as a test using characteristic values of three or more semiconductor devices.

[0012] In the semiconductor device characteristic value estimation system, the database preferably includes a first storage unit and a second storage unit. The first storage unit preferably has a function of storing a process list group. The second storage unit preferably has a function of storing a characteristic value group of the semiconductor device.

[0013] Furthermore, in the above-mentioned semiconductor element characteristic value estimation system, it is preferable that the processing unit has a first processing unit having a comparison function, a second processing unit having a testing function, a third processing unit having a regression analysis function, and a fourth processing unit having a function of estimating the semiconductor element characteristic value from the second process list.

[0014] In the above-described semiconductor element characteristic value estimation system, the characteristic value of the semiconductor element estimated by the processing unit is preferably one or more of threshold voltage, subthreshold swing value, on-current, and field-effect mobility.

[0015] Another aspect of the present invention includes a first step of inputting a first process list included in a first lot and characteristic values of a first semiconductor device fabricated according to the first process list; a second step of collecting second process lists from a group of process lists that have a certain degree of similarity to the first process list; a third step of testing the characteristic values of the first semiconductor device and the characteristic values of the second semiconductor device fabricated according to the second process list; and a third step of performing an analysis of variance of the characteristic values of the first plurality of semiconductor devices among the first plurality of semiconductor devices fabricated according to the first plurality of process lists included in the first lot, and comparing the first plurality of process lists to determine whether different processes exist in the first plurality of process lists. a fourth step of recording whether a parameter for each of the first plurality of process lists affects the characteristic values of the first plurality of semiconductor elements; a fifth step of collecting, from the group of process lists, a third process list having a certain degree of similarity or higher with respect to each of the first plurality of process lists; a sixth step of performing a regression analysis between parameters for processes that affect the characteristic values of the first plurality of semiconductor elements and the characteristic values of the third semiconductor elements fabricated according to the third process list; and a seventh step of estimating, from the second plurality of process lists included in the second lot, the characteristic values of the second plurality of semiconductor elements to be fabricated according to the second plurality of process lists before fabricating the second plurality of semiconductor elements.

[0016] In the above-mentioned method for estimating characteristic values of semiconductor elements, it is preferable that in the third step, testing is performed, and if there is a significant difference between the characteristic values of the first semiconductor element and the characteristic values of the second semiconductor element, information is output. [Effects of the Invention]

[0017] According to one aspect of the present invention, a method for estimating characteristic values of a semiconductor device to be prototyped can be provided. Furthermore, according to one aspect of the present invention, a system for estimating characteristic values of a semiconductor device to be prototyped can be provided. Furthermore, according to one aspect of the present invention, a system for automatically comparing the process and characteristic values of a semiconductor device prototyped this time with the process and characteristic values of semiconductor devices prototyped in the past can be provided.

[0018] Note that the effects of one embodiment of the present invention are not limited to the effects listed above. The effects listed above do not preclude the existence of other effects. Note that the other effects are effects not mentioned in this section, which will be described below. Effects not mentioned in this section can be derived by a person skilled in the art from the description in the specification, drawings, etc., and can be extracted as appropriate from these descriptions. Note that one embodiment of the present invention has at least one of the effects listed above and / or other effects. Therefore, one embodiment of the present invention may not have the effects listed above in some cases. [Brief explanation of the drawings]

[0019] [Figure 1] FIG. 1 is a flow chart showing an example of a method for estimating a characteristic value of a semiconductor device. [Figure 2] Fig. 2A is a diagram showing an example of a lot configuration, and Fig. 2B is a diagram for explaining characteristic values. [Figure 3] FIG. 3 is a diagram illustrating an example of a method for estimating a characteristic value of a semiconductor element. [Figure 4] FIG. 4 is a diagram illustrating an example of a method for estimating a characteristic value of a semiconductor element. [Figure 5] FIG. 5 is a diagram illustrating an example of a method for estimating a characteristic value of a semiconductor element. [Figure 6] FIG. 6 is a diagram illustrating an example of a method for estimating a characteristic value of a semiconductor element. [Figure 7] FIG. 7 is a diagram illustrating an example of a method for estimating a characteristic value of a semiconductor element. [Figure 8]FIG. 8 is a diagram illustrating an example of a method for estimating a characteristic value of a semiconductor element. [Figure 9] FIG. 9 is a diagram illustrating an example of a system configuration. [Figure 10] FIG. 10 is a diagram illustrating an example of a system configuration. [Figure 11] FIG. 11 is a diagram illustrating an example of a system configuration. [Figure 12] FIG. 12 is a diagram illustrating an example of a system configuration. [Figure 13] FIG. 13 is a diagram illustrating a computer device. DETAILED DESCRIPTION OF THE INVENTION

[0020] The embodiments will be described in detail with reference to the drawings. However, the present invention is not limited to the following description, and it will be readily understood by those skilled in the art that various changes can be made in form and detail without departing from the spirit and scope of the present invention. Therefore, the present invention should not be interpreted as being limited to the description of the embodiments shown below.

[0021] In the configuration of the invention described below, the same parts or parts having similar functions are denoted by the same reference numerals in different drawings, and repeated explanations thereof will be omitted. Furthermore, when referring to similar functions, the same hatch pattern may be used and no particular reference numeral may be assigned.

[0022] Furthermore, for ease of understanding, the position, size, range, etc. of each component shown in the drawings may not represent the actual position, size, range, etc. Therefore, the disclosed invention is not necessarily limited to the position, size, range, etc. disclosed in the drawings.

[0023] It should also be noted that the ordinal numbers "first," "second," and "third" used in this specification are used to avoid confusion of components and are not intended to limit the number.

[0024] (Embodiment) In this embodiment, a method for estimating a characteristic value of a semiconductor element (a characteristic value estimation method for a semiconductor element) and a system for estimating a characteristic value of a semiconductor element (a characteristic value estimation system for a semiconductor element), which are one embodiment of the present invention, will be described with reference to FIGS. 1 to 13.

[0025] <Flow> In this section, an example of a method for estimating characteristic values of a semiconductor device will be described with reference to FIGS. 1, 2A, and 2B.

[0026] FIG. 1 is a flow diagram showing an example of a method for estimating characteristic values of a semiconductor element. As shown in FIG. 1, the method for estimating characteristic values of a semiconductor element includes steps S001 to S006. Note that there are tasks to be performed by a user before step S001 and before step S006. As will be described later, there are cases where the tasks to be performed by a user before step S001 do not need to be performed. Note that the user described in this specification includes a user of one aspect of the present invention, a person implementing one aspect of the present invention, a person planning a lot, a person implementing a lot, etc.

[0027] First, the operations that the user must perform before step S001 will be described.

[0028] The user plans the first lot 11. The first lot 11 is a lot that is planned before step S001 shown in FIG.

[0029] [1st Lot 11] Here, the configuration of the first lot 11 will be described with reference to FIG. 2A.

[0030] A lot is the smallest unit of a product when producing the same type of product. It can also be considered the number of prototypes produced when prototyping a product. For example, when manufacturing a product with semiconductor elements, one lot can be produced to produce a product with one or more semiconductor elements.

[0031] One or more substrates are prepared for each lot. A process list for manufacturing a product having semiconductor elements is prepared for each substrate. The process list includes multiple processes set in the order of manufacturing the product having semiconductor elements, and processing conditions are specified for each process.

[0032] When producing a product having semiconductor elements, the process list prepared for each substrate within a single lot is the same. On the other hand, when prototyping semiconductor elements or products having semiconductor elements, some of the processes in the process list prepared for each substrate within a single lot may differ. This embodiment assumes the case of prototyping semiconductor elements. Furthermore, it is assumed that multiple semiconductor elements are fabricated on a substrate in accordance with the process list prepared for that substrate. Therefore, hereinafter, multiple semiconductor elements fabricated on a single substrate may be referred to as a semiconductor element group or simply as semiconductor elements.

[0033] Furthermore, planning a lot means creating a process list for each board in the lot.

[0034] Fig. 2A is a diagram showing an example of the configuration of a first lot 11. As shown in Fig. 2A, substrates 21_1 to 21_n (n is a natural number) are prepared in the first lot 11. Hereinafter, the substrates 21_1 to 21_n may be collectively referred to as substrates 21.

[0035] Each board is assigned an ID, which will be referred to as the board ID.

[0036] 2A, process lists 31_1 to 31_n are prepared for the substrates 21_1 to 21_n, respectively. Hereinafter, the process lists 31_1 to 31_n may be collectively referred to as the process list 31.

[0037] The process list is associated with the board ID. In other words, the process list may be read or written based on the board ID.

[0038] Each process list 31 includes a plurality of processes set in the order of the semiconductor device fabrication process. The processes for fabricating a semiconductor device include, for example, film formation, cleaning, resist application, exposure, development, processing, heat treatment, inspection, and substrate transfer. The number of processes in each of the process lists 31_1 to 31_n may be the same or different.

[0039] Each process may be assigned an ID different from the substrate ID. Here, the ID assigned to a process is referred to as a process ID.

[0040] Furthermore, processing conditions are specified for each process set in the process list 31. For example, processing conditions for a film formation process include the equipment, as well as set values for temperature, pressure, power, flow rate, and the like. The processing conditions for a film formation process affect the film thickness, film quality, and the like of the film formed by the film formation process. Therefore, they may affect the characteristic values of semiconductor elements. Of course, the processing conditions for processes other than film formation, the presence or absence of processes, the order of processes, and the like may also affect the characteristic values of semiconductor elements.

[0041] The processing conditions are selected from a set of parameters (major parameters, or simply parameters) prepared in advance. Note that the parameters can also be added later.

[0042] For example, the processing conditions for a process are selected and specified from a plurality of parameters prepared in advance, that is, parameters are specified for the process.

[0043] The above parameters may be assigned IDs different from the substrate ID and the process ID. Here, the IDs assigned to the parameters are referred to as parameter IDs.

[0044] Furthermore, when planning the first lot 11, the user designates one substrate from among the n substrates (substrates 21_1 to 21_n) prepared in the first lot 11. Hereinafter, the designated substrate will be referred to as reference substrate 21R. Note that the user does not have to designate the one substrate.

[0045] The above is the description of the configuration of the first lot 11.

[0046] Next, the user carries out the first lot 11. That is, the user manufactures semiconductor devices in accordance with the process list prepared for the first lot 11.

[0047] Next, the user measures the electrical characteristics of each of the fabricated semiconductor elements. For example, the electrical characteristics of the semiconductor elements can be measured using the Id-Vg characteristics, which evaluate the temperature characteristics or threshold voltage of the semiconductor elements.

[0048] Next, characteristic values are calculated from the measured electrical characteristics. These characteristic values include threshold voltage (Vth), subthreshold swing (S value), on-current (Ion), and field-effect mobility (μFE). Hereinafter, characteristic values calculated from the measurement results of the electrical characteristics of a semiconductor element may be referred to as semiconductor element characteristic values, or simply as characteristic values.

[0049] Characteristic values 61_1 to 61_n are calculated for the semiconductor elements fabricated on the substrates 21_1 to 21_n, respectively. Each of the characteristic values 61_1 to 61_n includes one or more of the above-mentioned characteristic values (Vth, S value, Ion, μFE, etc.). For example, as shown in FIG. 2B, the characteristic value 61_1 of the semiconductor element fabricated on the substrate 21_1 includes characteristic values 61_1(1) to 61_1(q) (q is a natural number). Furthermore, each of the characteristic values 61_1(1) to 61_1(q) includes characteristic values equal to or less than the number of semiconductor elements fabricated on the substrate 21_1. The same applies to the characteristic values 61_2 to 61_n, and the characteristic values 60_1 to 60_m described below. Hereinafter, the characteristic values (characteristic values 61_1 to 61_n) of the semiconductor elements fabricated on the substrate 21 may be collectively referred to as characteristic values 61.

[0050] The characteristic values are associated with the board ID. In other words, the reading and writing of the characteristic values may be performed based on the board ID.

[0051] The above is the work that the user performs before step S001. After the first lot 11 is executed and the characteristic value 61 is calculated, the user inputs the process list 31 and the characteristic value 61 of the semiconductor device. After input is complete, the process proceeds to step S001.

[0052] <<Step S001>> In step S001, the process list 31R prepared for the reference substrate 21R is compared with each of the process lists prepared for substrates included in a previously processed lot. Note that the process lists are compared based on the processes and processing conditions (parameters) set in the process lists. Hereinafter, the substrates included in the previously processed lot will be collectively referred to as a substrate group 20 (substrates 20_1 to 20_m (m is a natural number)). The process lists prepared for each of the substrate group 20 will be collectively referred to as a process list group 30 (process lists 30_1 to 30_m). The characteristic values of the semiconductor elements included in each of the substrate group 20 will be collectively referred to as a semiconductor element characteristic value group 60, or simply as a characteristic value group 60 (characteristic values 60_1 to 60_m). In other words, step S001 is a step of comparing the process list 31R with the process list group 30.

[0053] First, check whether each process matches or does not match. If an ID has been assigned to each process, you can simply check whether the process ID has been added, deleted, or changed. Alternatively, you can write out all the processes in text and check the differences in the strings. Note that processes that are not directly related to the shape or configuration of semiconductor elements, such as inspection and substrate transfer, can be excluded from the comparison in advance. This can reduce the time required for the comparison.

[0054] If the processes match, check whether the processing conditions (parameters) for that process match or not. If an ID is assigned to the parameter, check whether the parameter ID has been added, deleted, or changed. Alternatively, you can write all the parameters to text and check the differences in the strings.

[0055] To check whether the processes match or do not match, and whether the processing conditions (parameters) of the processes match or do not match, for example, a diff algorithm may be used.

[0056] By comparing the process list 31R with the group of process lists 30, process lists with a certain degree of similarity to the process list 31R can be collected from the group of process lists 30. In other words, step S001 is a process of collecting process lists with a certain degree of similarity to the process list 31R from the group of process lists 30. Note that if a certain number of process lists with a certain degree of similarity or more have been collected, process lists with the same degree of similarity (process lists that match the process list 31R) or process lists with a higher degree of similarity may be extracted from the certain number of collected process lists. Here, the process list collected in step S001 is referred to as process list 35. The process list 35 is also a process list with a certain degree of similarity or more to the process list 31R. In other words, step S001 can be rephrased as a process of acquiring the process list 35.

[0057] If the reference substrate 21R is not specified, all of the process lists 31 are compared with the process list group 30, and process lists having a certain degree of similarity or higher with respect to each of the process lists 31 are collected from the process list group 30. In this case, the process lists having a certain degree of similarity or higher with respect to each of the process lists 31 can be regarded as process lists 35.

[0058] If one or more process lists with a certain degree of similarity or higher, or one or more process lists with the same degree of similarity or higher degree of similarity are collected from the group of process lists 30, the process proceeds to step S002.

[0059] <<Step S002>> In step S002, verification is performed using the characteristic values of the semiconductor devices fabricated on the reference substrate 21R and the characteristic values of the semiconductor devices fabricated on the substrate associated with the process list 35 by the substrate ID. The process list and the characteristic values of the semiconductor devices fabricated on the substrate associated with the process list by the substrate ID are associated with each other by the substrate ID. Hereinafter, the characteristic value of the semiconductor device fabricated on the reference substrate 21R will be referred to as characteristic value 61R. Also, the characteristic value of the semiconductor device fabricated on the substrate having the collected process list may be simply referred to as the collected characteristic value. Also, the characteristic value of the semiconductor device fabricated on the substrate associated with the process list 35 by the substrate ID will be referred to as characteristic value 65. In other words, characteristic value 65 can be said to be the characteristic value collected in step S001.

[0060] The above test is performed using the characteristic value 61R and the characteristic value of a semiconductor device manufactured according to a process list with the same or highest similarity among the process lists 35. Note that the above test is not limited to this, and may be performed using the characteristic value 61R and the multiple characteristic values collected in step 001. Also, the above test is performed for each characteristic value.

[0061] The above test may be carried out using a t-test or the like.

[0062] If the test results in no significant difference between the characteristic value 61R and the characteristic value 65, the process proceeds to step S003. If a significant difference is determined between the characteristic value 61R and the characteristic value 65, there is a possibility that the first lot 11 has not been carried out correctly. Therefore, if a significant difference is determined, the user is notified to check whether the first lot 11 has been carried out correctly. After notifying the user, the process ends.

[0063] <<Step S003>> In step S003, an analysis of variance is performed on the characteristic values 61 to determine whether the process changes made in the first lot 11 have an effect on the characteristic values, and whether or not there is an effect is recorded. First, an analysis of variance (also called an ANOVA) of the characteristic values 61 is performed between the substrates 21 included in the first lot 11, and the process lists 31 are compared. The analysis of variance is performed for each characteristic value.

[0064] For the above analysis of variance, Type 2 ANOVA, Type 3 ANOVA, non-parametric tests (Kruskal-Wallis test or Friedman test), etc. are used.

[0065] Note that statistical analysis may be performed before the above analysis of variance. By performing statistical analysis, it is possible to appropriately select the method to be used for the analysis of variance. Examples of statistical analysis include outlier detection, testing for normality of distribution, and testing for homogeneity of variance. If the statistical analysis determines that the characteristic values 61 have no outliers, are normal in distribution, and have homogeneity of variance, it is recommended to select a parametric test for the above analysis of variance. It is also recommended to use Type 2 ANOVA as the parametric test. This can improve the accuracy of the analysis of variance.

[0066] Furthermore, if the statistical analysis determines that the characteristic values 61 contain outliers, that the distribution is not normal, or that the variances are not homogeneous, it is advisable to select a nonparametric test as the analysis of variance. This improves the accuracy of the analysis of variance. Note that if there is no correspondence between the characteristic values 61 of the substrates 21, the Kruskal-Wallis test is used as the nonparametric test.

[0067] To detect outliers, the Local Outlier Factor is used. To test the normality of distribution, the Shapiro-Wilk test and the Kolmogorov-Smirnov test are used. In particular, the Shapiro-Wilk test is used. To test the homogeneity of variance, the Levene test, the Bartlett test, the Hartley test, etc. are used. In particular, the Levene test is used.

[0068] By performing the above analysis of variance, it is possible to determine whether there is a significant difference in the characteristic values included in the characteristic values 61 (characteristic values 61_1 to 61_n) between the substrates 21 (substrates 21_1 to 21_n) included in the first lot 11.

[0069] Furthermore, by comparing the process lists 31 (process lists 31_1 to 31_n) between the substrates 21 (substrates 21_1 to 21_n) included in the first lot 11, it is possible to extract processes that differ between the substrates 21.

[0070] In comparing the process lists, as explained in <<Step S001>>, the processes and processing conditions (parameters) set in the process lists are compared. First, it is confirmed whether each process matches. If a process matches, it is confirmed whether the processing conditions (parameters) for that process match. To confirm whether a process matches or mismatches and whether the processing conditions (parameters) for the process match or mismatch, for example, a diff algorithm may be used.

[0071] If the above analysis of variance determines that there is a significant difference in one or more of the characteristic values 61 (characteristic values 61_1 to 61_n) between the substrates 21 (substrates 21_1 to 21_n) included in the first lot 11, the process that differs between the substrates 21 is recorded as having a significant effect on the characteristic value. Hereinafter, a process that is recorded as having a significant effect on the characteristic value may be simply referred to as an influential process. On the other hand, if it is determined that there is no significant difference in the characteristic value 61 between the substrates 21, the process that differs between the substrates 21 is recorded as not having a significant effect on the characteristic value.

[0072] Note that, among the characteristic values 61 (characteristic values 61_1 to 61_n), there may be characteristic values that are determined to have a significant difference and characteristic values that are determined to have no significant difference between the substrates 21 (substrates 21_1 to 21_n) included in the first lot 11. Therefore, the presence or absence of an influence on the characteristic values of the processes that differ between the substrates 21 may be recorded for each characteristic value.

[0073] After the above recording is completed, the process proceeds to step S004.

[0074] <<Step S004>> In step S004, all of the process lists 31 (process lists 31_1 to 31_n) prepared for the substrates 21 (substrates 21_1 to 21_n) included in the first lot 11 are compared with the process list group 30, and a process list that differs only in the affected processes is collected from the process list group 30 for each of the process lists 31. Here, the process list collected in step S004 is referred to as a process list 37. The process list 37 is also a process list that differs only in the affected processes for each of the process lists 31. In other words, step S004 can be rephrased as a process of acquiring the process list 37.

[0075] In comparing the process lists, as explained in <<Step S001>>, the processes and processing conditions (parameters) set in the process lists are compared. First, it is confirmed whether each process matches. If a process matches, it is confirmed whether the processing conditions (parameters) for that process match. To confirm whether a process matches or mismatches and whether the processing conditions (parameters) for the process match or mismatch, for example, a diff algorithm may be used.

[0076] If one or more process lists that differ only in the affected process are collected from the process list group 30 for each process list 31, the process proceeds to step S005. Hereinafter, the characteristic value of the semiconductor device fabricated on the substrate associated with the process list 37 by the substrate ID will be referred to as characteristic value 67. In other words, characteristic value 67 can be said to be the characteristic value collected in step S004.

[0077] <<Step S005>> In step S005, machine learning is performed based on the parameters for the affected processes and the characteristic value 67. For example, regression analysis is preferably used as the machine learning. In this case, step S005 can be said to be a step of performing regression analysis between the parameters for the affected processes and the characteristic value 67. By using regression analysis, it is possible to analyze the correlation between the parameters for the affected processes and the characteristic value 67. Therefore, it is possible to estimate the characteristic value of the semiconductor element.

[0078] Specifically, in step S005, a regression analysis is performed using the parameters for the affected process as explanatory variables and the characteristic value 67 as a response variable. For example, linear regression is performed using the least squares method on the parameters for the affected process and the characteristic value 67.

[0079] Before performing the regression analysis, a variance analysis of the characteristic values 67 may be performed to confirm that the differences in the parameters for the affected processes are significant with respect to the characteristic values. By confirming that the differences are significant, it can be determined that the accuracy of the characteristic values of the semiconductor device estimated based on the results of the regression analysis is high.

[0080] For the above analysis of variance, Type 2 ANOVA, Type 3 ANOVA, non-parametric tests (Kruskal-Wallis test or Friedman test), etc. are used.

[0081] Note that statistical analysis may be performed before the above analysis of variance. By performing statistical analysis, it is possible to appropriately select the method to be used for the analysis of variance. Examples of statistical analysis include outlier detection, testing for normality of distribution, and testing for homogeneity of variance. If the statistical analysis determines that the characteristic values 67 have no outliers, are normal in distribution, and have homogeneity of variance, it is recommended to select a parametric test for the above analysis of variance. It is also recommended to use Type 2 ANOVA as the parametric test. This can improve the accuracy of the analysis of variance.

[0082] Furthermore, if the above statistical analysis determines that the characteristic value 67 contains outliers, is not normally distributed, or does not have homogeneity of variance, it is advisable to select a nonparametric test as the above analysis of variance. This improves the accuracy of the analysis of variance. Note that if there is no correspondence between the characteristic values 67 of the process list 37 and the boards associated by board ID, the Kruskal-Wallis test is used as the nonparametric test.

[0083] To detect outliers, the Local Outlier Factor is used. To test the normality of distribution, the Shapiro-Wilk test and the Kolmogorov-Smirnov test are used. In particular, the Shapiro-Wilk test is used. To test the homogeneity of variance, the Levene test, the Bartlett test, the Hartley test, etc. are used. In particular, the Levene test is used.

[0084] If the above analysis of variance reveals that the difference in the parameter for the affected process is significant with respect to the characteristic value, then a regression analysis of the parameter and the characteristic value can be performed. By performing this regression analysis, the characteristic value of the semiconductor element can be estimated from the parameter.

[0085] Regression analysis can be performed using linear regression, ridge regression, Lasso regression, elastic net, k-nearest neighbor method, regression tree, random forest, support vector regression, neural network, etc.

[0086] Although the above example illustrates a method of performing analysis of variance before regression analysis, a correlation coefficient may be calculated by regression analysis instead of analysis of variance, and used as a criterion for determining whether to adopt an estimated value for a semiconductor element. The correlation coefficient may be Pearson's product-moment correlation coefficient, Spearman's rank correlation coefficient, Kendall's rank correlation coefficient, or the like.

[0087] From the above, it is possible to estimate the characteristic values of the semiconductor element from the affected process and the parameters for that process.

[0088] After step S005 is completed, the process list 31 and the characteristic value 61 are stored in the process list group 30 and the characteristic value group 60, respectively. This storage may be performed after step S006 is completed.

[0089] This concludes the explanation of step S005.

[0090] Next, as a task to be performed by the user before step S006, the user plans the second lot 12. Note that the second lot 12 does not necessarily have to be planned after step S005 is performed. The user may plan the second lot 12 before step S001 is started, or may plan it while steps S001 to S005 are being performed.

[0091] After the second lot 12 is planned, the process proceeds to step S006.

[0092] <<Step S006>> In step S006, characteristic value 62 is estimated from process list 32 based on the results of the regression analysis performed in step S005. Then, estimated characteristic value 62 is output. In other words, step S006 is a step of estimating characteristic value 62 and outputting estimated characteristic value 62. Here, process list 32 is a process list prepared for second lot 12. Furthermore, characteristic value 62 is a characteristic value of a semiconductor device to be manufactured in accordance with process list 32.

[0093] Note that what is output is not limited to the estimated characteristic value 62. For example, if the process lists 32 prepared for the substrates 22 included in the second lot 12 differ only in the non-influential processes, the user may be notified of information that there is a risk that the difference in characteristic value 62 between the substrates 22 may be small. Furthermore, the user may be notified of information regarding which processes in the process list 32R prepared for the reference substrate 22R specified in the second lot 12 affect the characteristic value 62.

[0094] After outputting the estimated characteristic value 62 or notifying the user of the above information, the process ends.

[0095] By performing steps S001 to S006, it is possible to estimate the characteristic values of a semiconductor element without fabricating the semiconductor element or measuring the electrical characteristics of the semiconductor element, thereby reducing the number of times that semiconductor elements are prototyped, thereby reducing development costs and shortening development time.

[0096] Furthermore, by using the method for estimating the characteristic values of semiconductor devices described in this section, data such as characteristic values, process lists, and process parameters are appropriately associated and stored. Therefore, as data is accumulated, the ability to estimate characteristic values improves, enabling accurate estimation of characteristic values. Furthermore, by planning the second lot 12, the process list 32 and the process list group 30 are automatically compared, allowing the user to analyze the data at a high level.

[0097] The above is a description of an example of a method for estimating the characteristic values of a semiconductor element.

[0098] <Example of flow> In this section, an example of steps S001 to S006 described in the previous <Flow> will be described with reference to FIGS.

[0099] FIG. 3 is a diagram illustrating an example of step S001.

[0100] In step S001, as described above, the process list 31R is compared with the process list group 30. For example, in FIG. 3, among the process lists 30_1 to 30_m, the process list 30_2 is assumed to be a process list having a certain degree of similarity with the process list 31R or higher. At this time, the process list 30_2 is collected. The process list 30_2 shown in FIG. 3 corresponds to the process list 35 described in the previous <Flow>. Then, the process proceeds to step S002.

[0101] FIG. 4 is a diagram illustrating an example of step S002.

[0102] In step S002, as described above, the characteristic value 61R and the characteristic value 65 are tested.

[0103] When a board ID is assigned, for example, in FIG. 4, a characteristic value 60_2 is extracted from the group of characteristic values 60 based on the board ID associated with the process list 30_2 collected in step S001. Note that the characteristic value 60_2 shown in FIG. 4 corresponds to the characteristic value 65 described in the previous <Flow>. Then, a test is performed using the characteristic value 61R (characteristic value 61R(1) to characteristic value 61R(p)) and the characteristic value 60_2 (characteristic value 60_2(1) to characteristic value 60_2(p)). If the test determines that there is no significant difference between the characteristic value 61R and the characteristic value 60_2, the process proceeds to step S003.

[0104] FIG. 5 is a diagram illustrating an example of step S003.

[0105] In step S003, as described above, an analysis of variance of the characteristic values 61 and a comparison of the process lists 31 are performed between the substrates 21. For example, in Fig. 5, by comparing the process lists 31 (process lists 31_1 to 31_n) between the substrates 21, it is determined that the process that differs between the substrates 21 is process A. Note that while Fig. 5 shows an example in which the process that differs between the substrates 21 is process A, the process that differs between the substrates 21 may be two or more processes.

[0106] 5, for example, a variance analysis of the characteristic values 61 (characteristic values 61_1 to 61_n) among the substrates 21 determines that there is a significant difference in the characteristic values included in the characteristic values 61. At this time, after process A is recorded as an affected process, the process proceeds to step S004.

[0107] FIG. 6 is a diagram illustrating an example of step S004.

[0108] In step S004, as described above, all process lists 31 are compared with the group of process lists 30, and for each process list 31, process lists that differ only in the affected process are collected from the group of process lists 30. For example, in FIG. 6, this comparison results in process lists 30_1 and 31_2 being collected as process lists that differ only in process A. Here, process lists 30_1 and 31_n correspond to the process list 37 described above in the <Flow>. Note that similarly, process lists that differ only in process A are collected for each of process lists 31_2 through 31_n. Then, the process proceeds to step S005.

[0109] FIG. 7 is a diagram illustrating an example of step S005.

[0110] In step S005, as described above, a regression analysis is performed between the parameters for the affected process and the characteristic value 67. In FIG. 7, the characteristic value 67 is a characteristic value of a semiconductor device fabricated on a substrate associated with the process list 30_1, etc., collected in step S004 by a substrate ID. That is, the characteristic value 67 is a characteristic value 60_1, etc. Also, for example, in FIG. 7, it is confirmed by analysis of variance that the difference in the parameters for process A in the process list 31_1 is significant with respect to the characteristic value. At this time, a regression analysis is performed between the parameters for process A and characteristic values such as characteristic value 61_1 and characteristic value 60_1. Then, the process proceeds to step S006.

[0111] FIG. 8 is a diagram illustrating an example of step S006.

[0112] As described above, in step S006, the characteristic values 62 are estimated from the process list 32 based on the results of the regression analysis performed in step S005. For example, in Fig. 8, the characteristic values 62 (characteristic values 62_1 to 62_n) are estimated from the process list 32 (process lists 32_1 to 32_n) based on the results of the regression analysis performed in step S005. Then, the estimated characteristic values 62 are output, and the process ends.

[0113] The above is an explanation of an example of steps S001 to S006.

[0114] <System configuration example> In this section, a configuration example of a system for estimating characteristic values of a semiconductor element, which is one embodiment of the present invention, will be described with reference to FIG.

[0115] Fig. 9 is a diagram showing an example of the configuration of a system 100 capable of estimating characteristic values of a semiconductor device. The system 100 includes an input unit (not shown in Fig. 9), a database 110, and a processing unit 120. The database 110 and the processing unit 120 are connected via a network. The network may include a local area network (LAN) or the Internet. The network may use either or both of wired and wireless communication.

[0116] The input section receives input of a process list 31, a process list 32, and a characteristic value 61.

[0117] The database 110 stores a group of process lists 30 prepared for the group of substrates 20 and a group of characteristic values 60 possessed by the group of substrates 20 .

[0118] The processing unit 120 includes a processing unit 120A and a processing unit 120B.

[0119] The processing unit 120A receives the process list 31 and the characteristic value 61 via the input unit. The processing unit 120A also receives the process list group 30 and the characteristic value group 60 stored in the database 110.

[0120] The processing unit 120A has a function of processing steps S001 to S005 described above. That is, the processing unit 120A has a function of comparing two process lists, a function of performing testing using characteristic values of two or more semiconductor devices, and a function of performing regression analysis between process parameters and characteristic values of two or more semiconductor devices. The testing includes analysis of variance. The processing unit 120A may also have a function of performing statistical analysis.

[0121] The two process lists are selected from a process list 31 input via the input unit and a group of process lists 30 stored in the database 110. The characteristic values of the two or more semiconductor elements are selected from a characteristic value 61 of the semiconductor element input via the input unit and a group of characteristic values 60 of the semiconductor element stored in the database 110. The characteristic values of the two or more semiconductor elements used for testing may not be the same as the characteristic values of the two or more semiconductor elements used for regression analysis.

[0122] Processing unit 120A may also have a function to output out1. Here, out1 is information for notifying the user to confirm whether the first lot has been executed correctly. By outputting this information, the user can know that the first lot may not have been executed correctly, without having to compare the first lot with past lots.

[0123] The process list 32 is input to the processing unit 120B via the input unit. The result of the regression analysis performed by the processing unit 120A is also input to the processing unit 120B.

[0124] The processing unit 120B has a function of processing the above-mentioned step S006. That is, the processing unit 120B has a function of estimating the characteristic value 62 from the process list 32 input via the input unit.

[0125] Furthermore, processing unit 120B has a function of outputting out2, where out2 is the estimated characteristic value 62 or the information explained in <<Step S006>>.

[0126] The above configuration provides a system 100 capable of estimating characteristic values of semiconductor devices. Furthermore, by using the system 100, data such as characteristic values, process lists, and process parameters are appropriately associated and stored. Therefore, as data is accumulated, the ability to estimate characteristic values improves, enabling accurate estimation of characteristic values. Furthermore, by planning the second lot 12, the process list 32 and the process list group 30 are automatically compared, allowing the user to analyze the data from a bird's-eye view.

[0127] <Details of the system configuration example> In this section, details of a configuration example of a system 100 according to one embodiment of the present invention will be described with reference to Fig. 10. Unless otherwise specified below, the components of the system described in this section, the functions of the components, and the like can be understood by referring to the description of the components of the system 100, the functions of the components, and the like.

[0128] FIG. 10 is a diagram showing a system 100A capable of estimating characteristic values of a semiconductor element. Note that the system 100A is a detailed configuration of the system 100 shown in FIG. 9. Like the system 100, the system 100A has an input unit (not shown in FIG. 10), a database 110, and a processing unit 120. Note that the database 110 has a memory unit 111 and a memory unit 112. Furthermore, the processing unit 120 has processing units 121 to 124.

[0129] <<Storage section 111>> The storage unit 111 stores a process list group 30. The storage unit 111 also stores a parameter set prepared in advance. The storage unit 111 may also record the processes that differ between the substrates 21 and whether or not they have an effect on the characteristic value 61, as described in <<Step S003>>.

[0130] <<Storage section 112>> The storage unit 112 stores a group of characteristic values 60 .

[0131] <<Processing unit 121>> The processing unit 121 receives the process list 31 and the process list group 30 as input.

[0132] The processing unit 121 has a function of comparing process lists. As described above, the process lists are preferably compared using the diff algorithm. Therefore, the diff algorithm is preferably stored in a memory unit (not shown in FIG. 10) included in the processing unit 121. Alternatively, if the diff algorithm is stored in the memory unit 111 or the like, the diff algorithm is preferably supplied to the processing unit 121 from the memory unit 111 or the like.

[0133] For example, the processing unit 121 can compare the process list 31R described in <<Step S001>> with the process list group 30. The processing unit 121 collects process lists from the process list group 30 that have a certain degree of similarity with the process list 31R. The processing unit 121 then outputs the collected process lists or the board IDs associated with the collected process lists to the processing unit 122. The process lists output to the processing unit 122 correspond to the process list 35 described in <<Step S001>>.

[0134] Furthermore, for example, the processing unit 121 can compare the process lists 31 between the substrates 21 included in the first lot 11, as described in <<Step S003>>. The processing unit 121 extracts processes that differ between the substrates 21. Then, the processing unit 121 outputs the processes that differ between the substrates 21, or the process IDs of those processes, to the processing unit 122.

[0135] Furthermore, for example, the processing unit 121 can compare the process list 31 with the process list group 30 described in <<Step S004>>. For each process list 31, the processing unit 121 collects process lists from the process list group 30 that differ only in the affected processes. Then, the processing unit 121 outputs the collected process list or the board ID associated with the collected process list to the processing unit 123. The process list output to the processing unit 123 corresponds to the process list 37 described in <<Step S004>>.

[0136] <<Processing Unit 122>> The characteristic value 61 is input to the processing unit 122. Also, the process list output from the processing unit 121, or the substrate ID associated with the process list, is input. Also, the process that differs between the substrates 21, or the process ID of the process, output from the processing unit 121, is input.

[0137] The processing unit 122 has a function of collecting characteristic values corresponding to the process list or the board ID from the characteristic value group 60. Note that the characteristic value group 60 may be input to the processing unit 122, and the characteristic values corresponding to the process list or the board ID may be extracted from the characteristic value group 60.

[0138] Furthermore, the processing unit 122 has a function of testing characteristic values. The testing includes analysis of variance. The processing unit 122 has a function of outputting out1. The processing unit 122 may also have a function of performing statistical analysis. Algorithms for analysis of variance, statistical analysis, etc. may be stored in a memory unit (not shown in FIG. 10) included in the processing unit 122. Alternatively, when algorithms for analysis of variance, statistical analysis, etc. are stored in the memory unit 112 or the like, the algorithms for analysis of variance, statistical analysis, etc. may be supplied to the processing unit 122 from the memory unit 112 or the like.

[0139] For example, the processing unit 122 can perform a test using the characteristic value 61R described in <<Step S002>> and the characteristic value collected in step S001. If the test results in a determination that there is a significant difference, the processing unit 122 outputs out1. The characteristic value collected in step S001 corresponds to the characteristic value 65 described in <<Step S002>>. Furthermore, out1 is information to be notified to the user, as described in <<Step S002>>.

[0140] Furthermore, if the test results in no significant difference, the processing unit 122 can perform the analysis of variance described in <<Step S003>>. At this time, the processing unit 122 has a function of outputting to the storage unit 111 whether or not there is an effect on the characteristic values of the processes that differ between the substrates 21. The processing unit 122 may also perform the statistical analysis described in <<Step S003>>.

[0141] <<Processing unit 123>> The characteristic value 61 is input to the processing unit 123. In addition, the process list output from the processing unit 121 or the board ID associated with the process list is input.

[0142] The processing unit 123 has a function of collecting characteristic values corresponding to the process list or the board ID from the characteristic value group 60. Note that the characteristic value group 60 may be input to the processing unit 123, and the characteristic values corresponding to the process list or the board ID may be extracted from the characteristic value group 60.

[0143] The processing unit 123 has a function of performing a regression analysis between the parameters for the affected processes and the characteristic values, as described in <<Step S005>>. The processing unit 123 also has a function of outputting the results of the regression analysis to the processing unit 124. The processing unit 123 also has a function of outputting the process list 31 to the storage unit 111. The processing unit 123 also has a function of outputting the characteristic values 61 to the storage unit 112. The regression analysis algorithm may be stored in a storage unit (not shown in FIG. 10) included in the processing unit 123. Alternatively, if the regression analysis algorithm is stored in the storage unit 112 or the like, the regression analysis algorithm may be supplied to the processing unit 123 from the storage unit 112 or the like.

[0144] The processing unit 123 may have a function to test characteristic values. Such testing includes analysis of variance. The processing unit 123 may also have a function to perform statistical analysis. If the processing unit 123 does not have the function to test characteristic values or the function to perform statistical analysis, the analysis of variance and statistical analysis described in <<Step S005>> may be performed by the processing unit 122. Note that arrows corresponding to input and output of data, commands, etc. between the processing units 122 and 123 are not shown in FIG. 10.

[0145] <<Processing unit 124>> The process list 32 is input to the processing unit 124. The result of the regression analysis output from the processing unit 123 is also input to the processing unit 124.

[0146] The processing unit 124 has a function of estimating characteristic values from the process list 32 using the results of the regression analysis. The processing unit 124 also has a function of outputting out2. The characteristic values estimated from the process list 32 correspond to the characteristic values 62 described in <<Step S006>>. Furthermore, out2 is the estimated characteristic values, information to be notified to the user, etc., described in <<Step S006>>.

[0147] It is preferable that the results of the regression analysis performed by processing unit 123 be held in a temporary storage area of processing unit 123 or processing unit 124 until the process list 32 is input to processing unit 124. Alternatively, the results of the regression analysis performed by processing unit 123 may be stored in storage unit 111 or the like, and input to processing unit 124 at the same time that the process list 32 is input to processing unit 124.

[0148] This configuration makes it possible to provide a system capable of estimating the characteristic values of semiconductor elements.

[0149] <System Variations> Note that the configuration example of the system for estimating the characteristic values of a semiconductor element is not limited to the configuration of the system 100A shown in Fig. 10. Below, modified examples of the system for estimating the characteristic values of a semiconductor element will be described with reference to Figs. 11 and 12.

[0150] <<System Variation 1>> Fig. 11 is a diagram showing an example of the configuration of a system 100B. In the system 100B shown in Fig. 11, elements having the same functions as elements constituting the systems shown in <System Configuration Example> and <Details of System Configuration Example> are denoted by the same reference numerals.

[0151] The system 100B shown in Fig. 11 is a modified example of the system 100A shown in Fig. 10. The system 100B differs from the system 100A in that it includes a storage unit 113.

[0152] The memory unit 113 stores the results of the regression analysis performed by the processing unit 123. This allows the system 100B to be in a standby state until the process list 32 is input.

[0153] <<System Variation 2>> Fig. 12 shows a configuration example of a system 100C. In the system 100C shown in Fig. 12, elements having the same functions as elements constituting the systems shown in <System Configuration Example> and <Details of System Configuration Example> are denoted by the same reference numerals.

[0154] A system 100C shown in Fig. 12 is a modified example of the system 100A shown in Fig. 10. The system 100C differs from the system 100A in that the process list 31 is stored in the memory unit 111 and the characteristic value 61 is stored in the memory unit 112 before performing the above-described steps S001 to S006.

[0155] With the above configuration, the process list and characteristic values of the executed lots are sequentially stored in the database, so that the system 100C can be used without planning and executing the first lot 11 before planning the second lot 12. In other words, the above-mentioned work that the user should do before step S001 does not need to be performed.

[0156] For example, a process list for a substrate to be used as a reference substrate is selected from the process list group 30. Next, process lists that have a certain degree of similarity to the selected process list are collected from the process list group 30 via the processing unit 121. The collected process lists are used as the process lists to be prepared for the first lot 11, and the selected process list is used as the process list 31R when using the system 100C. This allows the system 100C to be used without having to plan and implement the first lot 11.

[0157] <Computer equipment> In this section, a computer device including a system for estimating a characteristic value of a semiconductor element, which is one embodiment of the present invention, will be described with reference to FIG.

[0158] 13 is a diagram illustrating a computer having a system for estimating characteristic values of a semiconductor device. The computer 1000 is connected to a database 1011, a remote computer 1012, and a remote computer 1013 via a network. The computer 1000 includes an arithmetic unit 1001, a memory 1002, an input / output interface 1003, a communication device 1004, and a storage device 1005. The computer 1000 is electrically connected to a display device 1006a and a keyboard 1006b via the input / output interface 1003. The computer 1000 is also electrically connected to a network interface 1007 via the communication device 1004, and the network interface 1007 is electrically connected to the database 1011, the remote computer 1012, and the remote computer 1013 via a network.

[0159] Here, the network includes a local area network (LAN) and the Internet. The network can use either or both of wired and wireless communication. When wireless communication is used in the network, various communication methods can be used, such as short-range communication methods such as Wi-Fi (registered trademark) and Bluetooth (registered trademark), as well as communication methods conforming to the third generation mobile communication system (3G), LTE (sometimes called 3.9G), fourth generation mobile communication system (4G), or fifth generation mobile communication system (5G).

[0160] In the system for estimating characteristic values of a semiconductor element according to one aspect of the present invention, database 110 corresponds to database 1011. Note that database 110 may be storage 1005. Also, process list 31 and characteristic values 61 may be stored in storage 1005 first, and then stored in database 1011 after the completion of step S005 described above.

[0161] Furthermore, the processing unit 120 corresponds to the arithmetic device 1001. The processing unit 120 may be an arithmetic device included in the remote computer 1012 or the remote computer 1013. The processing units 121 to 123 included in the processing unit 120 may be arithmetic devices included in the remote computer 1012 or the remote computer 1013, and the processing unit 124 included in the processing unit 120 may be the arithmetic device 1001.

[0162] Furthermore, the display device 1006a displays the above-mentioned out1 and out2. Note that the above-mentioned results of the comparison of the process lists, the test results, the results of the analysis of variance, the results of the statistical analysis, etc. may be converted into, for example, a table, a mathematical formula, a graph, etc. and displayed on the display device 1006a.

[0163] As described above, one aspect of the present invention can provide a method for estimating characteristic values of a semiconductor device to be prototyped. Another aspect of the present invention can provide a system for estimating characteristic values of a semiconductor device to be prototyped. Another aspect of the present invention can provide a system for automatically comparing the process and characteristic values of a currently prototyped semiconductor device with the process and characteristic values of a previously prototyped semiconductor device.

[0164] This embodiment can be carried out by combining parts thereof as appropriate. [Explanation of symbols]

[0165] :11: First lot, 12: Second lot, 20: Group of substrates, 20_m: Substrate, 20_1: Substrate, 21: Substrate, 21_n: Substrate, 21_1: Substrate, 21R: Reference substrate, 22: Substrate, 22R: Reference substrate, 30: Process list group, 30_m: Process list, 30_1: Process list, 30_2: Process list, 31: Process list, 31_n: Process list, 31_1: Process list t, 31_2: Process list, 31R: Process list, 32: Process list, 32_n: Process list, 32_1: Process list, 32R: Process list, 35: Process list, 37: Process list, 60: Characteristic value group, 60_m: Characteristic value, 60_1: Characteristic value, 60_2: Characteristic value, 61: Characteristic value, 61_n: Characteristic value, 61_1: Characteristic value, 61_2: Characteristic value, 61R: Characteristic value, 62: Characteristic value property value, 62_n: property value, 62_1: property value, 65: property value, 67: property value, 100: system, 100A: system, 100B: system, 100C: system, 110: database, 111: memory unit, 112: memory unit, 113: memory unit, 120: processing unit, 120A: processing unit, 120B: processing unit, 121: processing unit, 122: processing unit, 123: processing unit, 124: processing unit, 1000: computer device, 1001: arithmetic unit, 1002: memory, 1003: input / output interface, 1004: communication device, 1005: storage, 1006a: display device, 1006b: keyboard, 1007: network interface, 1011: database, 1012: remote computer, 1013: remote computer

Claims

1. An input unit, a database, a processing unit, and a first storage unit, a first process list, a second process list, and characteristic values of a semiconductor device are input to the input unit; the database has a function of storing a group of process lists and a group of characteristic values of semiconductor devices; The processing unit a function of comparing two process lists selected from the first process list and the group of process lists; a function of performing testing using the characteristic values of the semiconductor device and two or more characteristic values of the semiconductor device selected from the group of characteristic values of the semiconductor device; a function of performing a regression analysis between process parameters and the characteristic values of the semiconductor device and two or more characteristic values of the semiconductor device selected from the group of characteristic values of the semiconductor device; a function of estimating characteristic values of a semiconductor device from the second process list; The first storage unit has a function of storing the results of the regression analysis performed by the processing unit.

2. In claim 1, For the comparison, the diff algorithm is used, Among the above tests, A t-test is used as a test using the characteristic values of two semiconductor elements. A characteristic value estimation system for semiconductor devices, which uses a non-parametric test as a test using characteristic values of three or more semiconductor devices.

3. In claim 1 or claim 2, the database has a second storage unit and a third storage unit, the second storage unit has a function of storing the process list group, The characteristic value estimation system for a semiconductor device, wherein the third storage unit has a function of storing a group of characteristic values of the semiconductor device.

4. In any one of claims 1 to 3, The processing unit a first processing unit having a function of performing the comparison; a second processing unit having a function of performing the assay; a third processing unit having a function of performing the regression analysis; a fourth processing unit having a function of estimating a characteristic value of a semiconductor element from the second process list.

5. In any one of claims 1 to 4, The characteristic value estimation system for a semiconductor element, wherein the characteristic value of the semiconductor element estimated by the processing unit is one or more of a threshold voltage, a subthreshold swing value, an on-current, and a field-effect mobility.

Images