Semiconductor device characteristic value estimation system

Abstract

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

Description

Technical Field

[0001] One aspect of the present invention relates to a method for estimating characteristic values of a semiconductor device. Another aspect of the present invention relates to a system for estimating characteristic values of a semiconductor device. Note that, in this specification and the like, a semiconductor device refers to a device that can function by utilizing semiconductor characteristics. As an example, it is a semiconductor device such as a transistor, a diode, a light-emitting device, or a light-receiving device. Another example of a semiconductor device is a passive device generated by a conductive film or an insulating film such as a capacitor, a resistor, or an inductor. Another example of a semiconductor device is a semiconductor device having a semiconductor device or a circuit including a passive device.

[0002] In addition, in this specification and the like, a semiconductor device refers to a device that can function by utilizing semiconductor characteristics. As an example, it is a semiconductor device such as a transistor, a diode, a light-emitting device, or a light-receiving device. Another example of a semiconductor device is a passive device generated by a conductive film or an insulating film such as a capacitor, a resistor, or an inductor. Another example of a semiconductor device is a semiconductor device having a semiconductor device or a circuit including a passive device. As an example, it is a semiconductor device such as a transistor, a diode, a light-emitting device, or a light-receiving device. Another example of a semiconductor device is a passive device generated by a conductive film or an insulating film such as a capacitor, a resistor, or an inductor. Another example of a semiconductor device is a semiconductor device having a semiconductor device or a circuit including a passive device. As an example, it is a semiconductor device such as a transistor, a diode, a light-emitting device, or a light-receiving device. Another example of a semiconductor device is a passive device generated by a conductive film or an insulating film such as a capacitor, a resistor, or an inductor. Another example of a semiconductor device is a semiconductor device having a semiconductor device or a circuit including a passive device. As an example, it is a semiconductor device such as a transistor, a diode, a light-emitting device, or a light-receiving device. Another example of a semiconductor device is a passive device generated by a conductive film or an insulating film such as a capacitor, a resistor, or an inductor. Another example of a semiconductor device is a semiconductor device having a semiconductor device or a circuit including a passive device. As an example, it is a semiconductor device such as a transistor, a diode, a light-emitting device, or a light-receiving device. Another example of a semiconductor device is a passive device generated by a conductive film or an insulating film such as a capacitor, a resistor, or an inductor. Another example of a semiconductor device is a semiconductor device having a semiconductor device or a circuit including a passive device.

Background Art

[0003] In recent years, in the fields using artificial intelligence (AI), the field of robots, or the energy field handling high power such as power ICs, development of new semiconductor devices has been underway to solve problems such as an increase in the amount of calculation or an increase in power consumption. The integrated circuits or semiconductor devices used in integrated circuits required by the market are becoming more complex, and an early startup of integrated circuits having new functions is required. However, in process design, device design, or circuit design in the development of semiconductor devices, knowledge, know-how, or experience of skilled engineers is required. In recent years, in the fields using artificial intelligence (AI), the field of robots, or the energy field handling high power such as power ICs, development of new semiconductor devices has been underway to solve problems such as an increase in the amount of calculation or an increase in power consumption. The integrated circuits or semiconductor devices used in integrated circuits required by the market are becoming more complex, and an early startup of integrated circuits having new functions is required. However, in process design, device design, or circuit design in the development of semiconductor devices, knowledge, know-how, or experience of skilled engineers is required. In recent years, in the fields using artificial intelligence (AI), the field of robots, or the energy field handling high power such as power ICs, development of new semiconductor devices has been underway to solve problems such as an increase in the amount of calculation or an increase in power consumption. The integrated circuits or semiconductor devices used in integrated circuits required by the market are becoming more complex, and an early startup of integrated circuits having new functions is required. However, in process design, device design, or circuit design in the development of semiconductor devices, knowledge, know-how, or experience of skilled engineers is required. In recent years, in the fields using artificial intelligence (AI), the field of robots, or the energy field handling high power such as power ICs, development of new semiconductor devices has been underway to solve problems such as an increase in the amount of calculation or an increase in power consumption. The integrated circuits or semiconductor devices used in integrated circuits required by the market are becoming more complex, and an early startup of integrated circuits having new functions is required. However, in process design, device design, or circuit design in the development of semiconductor devices, knowledge, know-how, or experience of skilled engineers is required. In recent years, in the fields using artificial intelligence (AI), the field of robots, or the energy field handling high power such as power ICs, development of new semiconductor devices has been underway to solve problems such as an increase in the amount of calculation or an increase in power consumption. The integrated circuits or semiconductor devices used in integrated circuits required by the market are becoming more complex, and an early startup of integrated circuits having new functions is required. However, in process design, device design, or circuit design in the development of semiconductor devices, knowledge, know-how, or experience of skilled engineers is required. In recent years, in the fields using artificial intelligence (AI), the field of robots, or the energy field handling high power such as power ICs, development of new semiconductor devices has been underway to solve problems such as an increase in the amount of calculation or an increase in power consumption. The integrated circuits or semiconductor devices used in integrated circuits required by the market are becoming more complex, and an early startup of integrated circuits having new functions is required. However, in process design, device design, or circuit design in the development of semiconductor devices, knowledge, know-how, or experience of skilled engineers is required. In recent years, in the fields using artificial intelligence (AI), the field of robots, or the energy field handling high power such as power ICs, development of new semiconductor devices has been underway to solve problems such as an increase in the amount of calculation or an increase in power consumption. The integrated circuits or semiconductor devices used in integrated circuits required by the market are becoming more complex, and an early startup of integrated circuits having new functions is required. However, in process design, device design, or circuit design in the development of semiconductor devices, knowledge, know-how, or experience of skilled engineers is required.

[0004] In recent years, regarding semiconductor devices, methods for optimizing the manufacturing process, methods for estimating device characteristics, and the like have been proposed. In Patent Document 1, the cross-sectional shape of a semiconductor device In recent years, regarding semiconductor devices, methods for optimizing the manufacturing process, methods for estimating device characteristics, and the like have been proposed. In Patent Document 1, the cross-sectional shape of a semiconductor device Image features are calculated from the SEM images of the pattern, and the relationship between these image features and device characteristics is examined. A method for estimating the device characteristics of a pattern under evaluation from its response relationship is disclosed. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2007-129059 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] In the semiconductor device manufacturing process, there are many steps involved in completing a semiconductor device, and the types of steps The types and processing conditions are also diverse. Therefore, the characteristics calculated from the electrical characteristics of semiconductor devices This involves verifying the process dependence of values ​​(sometimes simply referred to as characteristic values ​​of semiconductor devices). This is difficult. Furthermore, the process and characteristic values ​​of the newly fabricated semiconductor device and those of previously fabricated devices are difficult to compare. Comparing the process and characteristic values ​​of semiconductor devices requires an enormous amount of effort.

[0007] Conventional technology involves a lengthy semiconductor device manufacturing process, which is based on previously prototyped semiconductors. It is difficult to compare with the manufacturing process of the device. Also, regarding newly prototyped semiconductor devices... Even if they exist, it is difficult to determine which process is unique.

[0008] Therefore, one aspect of the present invention provides a method for estimating the characteristic values ​​of a semiconductor device that is planned to be prototyped. One of the objectives is to achieve the following. Furthermore, one aspect of the present invention relates to the characteristics of a semiconductor device that is planned to be prototyped. One of the objectives is to provide a system for estimating values. Furthermore, one aspect of the present invention is described here. Process and characteristic values ​​of the prototype semiconductor device, and process and characteristics of semiconductor devices prototyped in the past. One of the objectives is to provide a system that automatically performs comparisons with values.

[0009] Furthermore, the description of these problems does not preclude the existence of other problems. One approach does not require that all of these issues be resolved. The title will become clear from the description in the specification, drawings, claims, etc. It is possible to extract other issues from the descriptions in the drawings, claims, etc. [Means for solving the problem]

[0010] One aspect of the present invention relates to the characteristics of a semiconductor element having an input unit, a database, and a processing unit. This is a value estimation system. The input section includes a first process list, a second process list, and semiconductor The characteristic values ​​of the semiconductor element are entered. The database contains a list of process steps and the characteristics of the semiconductor element. It has the function of storing a group of values. The processing unit has a first process list and a group of process lists. A function to compare two selected process lists, the characteristic values ​​of semiconductor devices, and semiconductors. A device that performs verification using the characteristic values ​​of two or more semiconductor devices selected from a set of characteristic values ​​for the devices. The parameters for the process, the characteristic values ​​of the semiconductor device, and the group of characteristic values ​​of the semiconductor device. A function to perform regression analysis with the characteristic values ​​of two or more semiconductor elements selected from among them, and the second process It has a function to estimate the characteristic values ​​of semiconductor devices from a list.

[0011] In the above semiconductor device characteristic value estimation system, the diff algorithm is used for comparison. It is preferable to use the following. In addition, among the tests, the test that uses the characteristic values ​​of two semiconductor elements and Then, as a test using t-test and using characteristic values of three or more semiconductor devices, it is preferable to use a nonparametric Kruskal-Wallis test.

[0012] Further, in the characteristic value estimation system of the semiconductor device, it is preferable that the database has a first storage unit and a second storage unit. Further, it is preferable that the first storage unit has a function of storing a process list group Further, it is preferable that the second storage unit has a function of storing a characteristic value group of semiconductor devices.

[0013] Further, in the characteristic value estimation system of the semiconductor device, the processing unit has a first processing unit having a function of performing comparison a second processing unit having a function of performing a test, a third processing unit having a function of performing a regression analysis and a fourth processing unit having a function of estimating a characteristic value of a semiconductor device from a second process list. It is preferable to have.

[0014] Further, in the characteristic value estimation system of the semiconductor device, the characteristic value of the semiconductor device estimated by the processing unit is preferably at least one of a threshold voltage, a subthreshold swing value, an on-current, and a field-effect mobility.

[0015] Another aspect of the present invention is a first step of inputting a first process list included in a first lot and characteristic values of a first semiconductor device manufactured according to the first process list, and a second step of collecting a second process list having a similarity of a certain level or more with respect to the first process list from a process list group, and a third step of performing a test using the characteristic values of the first semiconductor device and the characteristic values of a second semiconductor device manufactured according to the second process list, and a fourth step of estimating a characteristic value of a semiconductor device from the second process list, and a fifth step of manufacturing a second semiconductor device according to the second process list, and ​Between the first multiple semiconductor elements, which are fabricated according to the first multiple process list included in the to Analysis of variance of the characteristic values ​​of the first set of semiconductor elements, and the ratio of the first set of process lists. A comparison is made, and different processes in the first list of multiple processes result in the characteristic values ​​of the first list of multiple semiconductor devices. The fourth step is to record whether or not it has an impact, and each of the first multiple process lists. In contrast, a third process list with a similarity above a certain level is collected from the group of process lists in a fifth process. The steps and parameters for the process that affect the characteristic values ​​of the first set of semiconductor elements Then, a regression analysis was performed with the characteristic values ​​of the third semiconductor device, which was manufactured according to the third process list. The sixth step and the second multiple process list included in the second lot, The characteristic values ​​of the second set of semiconductor devices manufactured according to the process list are... A method for estimating the characteristic values ​​of a semiconductor device, comprising a seventh step of estimation before fabricating the device. It is the law.

[0016] In the above semiconductor device characteristic value estimation method, the third step is to perform a test, and the first If there is a significant difference between the characteristic values ​​of one semiconductor device and the characteristic values ​​of a second semiconductor device, information will be released. It is preferable to be forced. [Effects of the Invention]

[0017] According to one aspect of the present invention, a method for estimating the characteristic values ​​of a semiconductor device to be prototyped is provided. Furthermore, according to one aspect of the present invention, the characteristic values ​​of a semiconductor device to be prototyped can be estimated. A system can be provided that does this. Furthermore, according to one aspect of the present invention, the semiconductor prototyped this time A comparison of the process and characteristic values ​​of the body element with the process and characteristic values ​​of semiconductor elements previously prototyped. We can provide a system that performs the task automatically.

[0018] The effects of one embodiment of the present invention are not limited to those listed above. This does not preclude the existence of other effects. These other effects are described below. This is an effect not mentioned in the section. An effect not mentioned in this section would be obvious to someone skilled in the art. This can be derived from detailed descriptions, drawings, etc., and can be extracted appropriately from these descriptions. Yes, it is possible. Furthermore, one aspect of the present invention includes, among the effects listed above and / or other effects, at least one of the following: It has at least one effect. Therefore, one aspect of the present invention may, in some cases, It may not always have the effects listed above. [Brief explanation of the drawing]

[0019] [Figure 1] Figure 1 is a flowchart illustrating an example of a method for estimating the characteristic values ​​of a semiconductor device. [Figure 2] Figure 2A shows an example of lot configuration. Figure 2B is a diagram illustrating characteristic values. [Figure 3] Figure 3 illustrates an example of a method for estimating the characteristic values ​​of a semiconductor device. [Figure 4] Figure 4 illustrates an example of a method for estimating the characteristic values ​​of a semiconductor device. [Figure 5] Figure 5 illustrates an example of a method for estimating the characteristic values ​​of a semiconductor device. [Figure 6] Figure 6 illustrates an example of a method for estimating the characteristic values ​​of a semiconductor device. [Figure 7] Figure 7 illustrates an example of a method for estimating the characteristic values ​​of a semiconductor device. [Figure 8] Figure 8 illustrates an example of a method for estimating the characteristic values ​​of a semiconductor device. [Figure 9] Figure 9 shows an example of the system configuration. [Figure 10]Figure 10 shows an example of the system configuration. [Figure 11] Figure 11 shows an example of the system configuration. [Figure 12] Figure 12 shows an example of the system configuration. [Figure 13] Figure 13 is a diagram illustrating a computer device. [Modes for carrying out the invention]

[0020] Embodiments will be described in detail with reference to the drawings. However, the present invention is not limited to the following description. Without departing from the spirit and scope of the present invention, its form and details may be changed in various ways. Those skilled in the art will readily understand that further improvements are possible. Therefore, the present invention can be implemented as follows: It is not to be interpreted as being limited to the description of the form.

[0021] Furthermore, in the configuration of the invention described below, the same part or part having a similar function The same reference numeral is used in common across different drawings, and explanations of its repetition are omitted. When referring to a specific function, the hatch pattern may be the same, and no special designation may be assigned.

[0022] Furthermore, the position, size, and scope of each component shown in the drawings are, for the sake of ease of understanding, actual The position, size, and range of the edges may not be shown. Therefore, the disclosed invention is not necessarily However, this is not limited to the location, size, and scope disclosed in the drawings.

[0023] Furthermore, the ordinal numbers "1st," "2nd," and "3rd" used in this specification refer to the constituent elements. This note is added to avoid confusion and does not imply any numerical limitation.

[0024] (Embodiment) In this embodiment, using Figures 1 to 13, the characteristics of a semiconductor device, which is one aspect of the present invention, are described. Method for estimating characteristic values ​​(method for estimating characteristic values ​​of semiconductor devices), and method for estimating characteristic values ​​of semiconductor devices This section describes a system (a system for estimating the characteristic values ​​of semiconductor devices).

[0025] <flow> This section presents an example of a method for estimating the characteristic values ​​of semiconductor devices, as shown in Figures 1, 2A, and 2B. We will explain using this method.

[0026] Figure 1 is a flowchart illustrating an example of a method for estimating the characteristic values ​​of a semiconductor device. As described above, the method for estimating the characteristic values ​​of a semiconductor device is as follows: Steps S001 to S006 It has. Furthermore, up to step S001 and up to step S006, it has Each of these tasks requires user intervention. As will be explained later, the user must perform the following tasks before step S001. The work may not be required to be performed. Furthermore, users described in this specification are entitled to this publication. Users of one aspect of the present invention, implementers of one aspect of the present invention, lot planners, lot implementers, etc. It is included.

[0027] First, we will explain the tasks that the user must perform before step S001.

[0028] The user plans the first lot 11. Note that the first lot 11 is as shown in Figure 1. This is a batch planned before the S001 model.

[0029] [Lot 1, No. 11] Here, the configuration of the first lot 11 will be explained using Figure 2A.

[0030] A lot refers to the smallest unit of production of the same type of product. This can be considered the number of prototypes produced when creating a prototype. For example, when manufacturing a product that has semiconductor elements. In this case, by performing one lot, it is possible to manufacture a product having one or more semiconductor elements. It is possible.

[0031] Each lot contains one or more substrates. Furthermore, each substrate contains a semiconductor. A process list is prepared for manufacturing a product containing the element. Furthermore, the process list includes multiple steps. The number of steps is set in the order of the manufacturing process for products containing semiconductor elements, and processing conditions are set for each step. This is specified.

[0032] Furthermore, when producing products containing semiconductor elements, within a single lot, each substrate is prepared The process list is the same. On the other hand, when testing semiconductor elements or products containing semiconductor elements... When manufacturing, within a single lot, the process list prepared for each substrate may include a portion of the process. There may be differences. In this embodiment, it is assumed that a semiconductor device is being prototyped. Furthermore, multiple semiconductor elements are fabricated on the substrate according to the process list provided for that substrate. It shall be manufactured. Therefore, from here on, multiple semiconductor elements fabricated on a single substrate will be used. It may also be referred to as a group of semiconductor devices, or simply as semiconductor devices.

[0033] Furthermore, lot planning involves creating a process list for each circuit board within that lot. It refers to accomplishing something.

[0034] Figure 2A shows an example of the configuration of the first lot 11. As shown in Figure 2A, the first lot Set 11 is provided with circuit boards 21_1 through 21_n (where n is a natural number). In some cases, boards 21_1 through 21_n are collectively referred to as board 21.

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

[0036] Furthermore, as shown in Figure 2A, each of the substrates 21_1 to 21_n has a process lithography Process list 31_1 to process list 31_n are prepared. From here on, process list 31_1 to process list 31_n are prepared. The process list 31_n is sometimes referred to collectively as process list 31.

[0037] Note that the process list is associated with the substrate ID. In other words, the process list can be read. In some cases, the initial output may be based on the circuit board ID.

[0038] Furthermore, each step in the process list 31 contains multiple steps, arranged in the order of the semiconductor device manufacturing process. The process for manufacturing semiconductor devices includes, for example, film deposition, cleaning, resist coating, dew These include light, development, processing, heat treatment, inspection, and substrate transfer. (See process list 31_1 to...) The number of steps in each step of process list 31_n may be the same or different.

[0039] Note that each process may be assigned an ID different from the substrate ID. The ID assigned to a process is referred to as the process ID.

[0040] Furthermore, processing conditions are specified for each of the processes set in process list 31. For example, the processing conditions for the film deposition process include the apparatus, as well as temperature, pressure, power, and flow rate. There are various setting values, etc. Note that the processing conditions for the film deposition process are determined by the film deposition process. This affects the film thickness, film quality, etc. Therefore, it may affect the characteristic values ​​of semiconductor devices. Yes, it is. Of course, it can also be half the amount depending on the processing conditions of steps other than film formation, the presence or absence of other steps, the order of steps, etc. This may affect the characteristic values ​​of the conductive element.

[0041] The above processing conditions are based on a pre-prepared set of parameters (main parameters, or It shall be selected from among the following parameters. You can also add it from there.

[0042] For example, the processing conditions for a process can be selected from a set of pre-defined parameters. Select and specify. In other words, specify parameters for the process.

[0043] Note that each of the above parameters has an ID that is different from the substrate ID and process ID. It is acceptable if they are assigned. Here, the ID assigned to the parameter is referred to as the parameter ID. To record.

[0044] Furthermore, when planning the first lot 11, the user will select the n provided within the first lot 11. From among the individual boards (boards 21_1 to boards 21_n), one board is selected. The specified board will be referred to as reference board 21R. The user must specify the single board. You don't have to.

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

[0046] Next, the user performs the first lot 11. That is, the first lot 11 is prepared The semiconductor device is manufactured according to the process list.

[0047] Next, the user measures the electrical characteristics of each semiconductor device they have fabricated. For example, semiconductor The electrical characteristics of a device are evaluated using the Id-Vg characteristics, which include the temperature characteristics or threshold voltage of the semiconductor device. You can use it.

[0048] Next, characteristic values ​​are calculated from the measured electrical characteristics. These characteristic values ​​include the threshold voltage (Vth). ), subthreshold swing value (S value), on-current (Ion), field effect mobility (μ Examples include FE). In the following sections, characteristic values ​​calculated from the measurement results of the electrical characteristics of semiconductor devices are described. These are sometimes called characteristic values ​​of semiconductor devices, or simply characteristic values.

[0049] For semiconductor devices fabricated on substrates 21_1 to 21_n, the characteristic value 6 The characteristic values ​​1_1 to 61_n are calculated. Each of these has one or more of the characteristic values ​​mentioned above (Vth, S value, Ion, μFE, etc.). Numbers are included. For example, as shown in Figure 2B, semiconductor elements fabricated on substrate 21_1 The characteristic value 61_1 includes characteristic value 61_1(1) to characteristic value 61_1(q) (where q is a natural number). It is included. Furthermore, each of the characteristic values ​​61_1(1) to 61_1(q) This is equal to, or less than, the number of semiconductor elements fabricated on substrate 21_1. This includes numerical characteristic values. Note that characteristic values ​​61_2 through 61_n, and the characteristics described below, are included. The same applies to values ​​60_1 and characteristic values ​​60_m, etc. Hereafter, the fabrication will be carried out on the substrate 21. The characteristic values ​​of the semiconductor device (characteristic values ​​61_1 to 61_n) are grouped together and given characteristic value 6 It is sometimes written as 1.

[0050] Note that characteristic values ​​are associated with the board ID. In other words, reading and writing characteristic values These procedures may be performed based on the circuit board ID.

[0051] The above is the work that the user must perform before step S001. The first lot 11 is actually After the process is completed and the characteristic value 61 is calculated, the user can access the process list 31 and the semiconductor device characteristics Enter the gender value 61. After the input is complete, proceed to step S001.

[0052] <<Step S001>> In step S001, the process list 31R prepared on the reference substrate 21R and previously performed A comparison is made between each of the process lists prepared for the substrates included in the lot and the following: In comparing process lists, the processes and processing conditions (parameters) set in the process list are This will be used for comparison. From here on, we will group together the circuit boards included in previously produced lots and refer to them as circuit boards. This group is denoted as group 20 (substrate 20_1 to substrate 20_m (where m is a natural number)). The process lists prepared for each of the board group 20 are combined into a process list group 30 (process list This will be denoted as 30_1 to process list 30_m). Also, each of the substrate group 20 is included The characteristic values ​​of semiconductor devices are grouped together as semiconductor device characteristic value group 60, or simply characteristic value group 6. 0 (characteristic value 60_1 to characteristic value 60_m) is represented. In other words, step S001 is, This process involves comparing the process list 31R with the process list group 30.

[0053] First, check for matches / mismatches in each process. Note that if an ID is assigned to a process, You just need to check if you've added, deleted, or changed the process ID. Or, enter all the processes into text. You can also write it out and check the difference in the strings. Note that the shape of semiconductor elements can be checked during inspection and board transfer. Processes that do not directly involve the structure, etc., can be excluded from the above comparison beforehand. This reduces the time required for the above comparison.

[0054] If the processes match, check whether the processing conditions (parameters) for those processes match or not. If an ID is assigned to a parameter, you can add, delete, or You can check the changes. Alternatively, write all the parameters to a text file and check the difference between the strings. You can investigate that.

[0055] For checking the consistency / mismatch of processes, and the consistency / mismatch of process processing conditions (parameters) For example, the diff algorithm can be used.

[0056] By comparing process list 31R with process list group 30, the following can be observed: Process lists with a similarity level above a certain threshold can be collected from the group of process lists 30. In other words, step S001 is a process that has a similarity to process list 31R above a certain level. This is a process of collecting processes from the process list group 30. Note that processes with a similarity level above a certain level If a certain number of process lists are collected, process lists with the same similarity (process list 31) A certain number of process lists (which match R) or have a higher similarity are collected. The process list collected in step S001 may be extracted from this list. The process list is denoted as process list 35. Process list 35 is similar to process list 31R. This is also a process list of a certain number of steps. In other words, step S001 obtains process list 35. This can be rephrased as a process that involves doing so.

[0057] If reference substrate 21R is not specified, the process will be applied to all of the process list 31. A comparison is made with list group 30, and for each of the process list 31, if the similarity is above a certain level... It is good practice to collect a certain process list from the process list group 30. At this time, process list 3 For each of the items in 1, the process list whose similarity is above a certain level is considered to be process list 35. It is possible.

[0058] A list of processes with a similarity level above a certain threshold, or processes with the same similarity level, or processes with the same similarity level. If one or more higher-level process lists are collected from the process list group 30, then step S Proceed to 002.

[0059] <<Step S002>> In step S002, the characteristic values ​​of the semiconductor device fabricated on the reference substrate 21R and the process rate are determined. The characteristic values ​​of the semiconductor device fabricated on the substrate associated with the STR35 and substrate ID, and the inspection The process is determined. The process list and the process on the board associated with the process list and the board ID are then used. The characteristic values ​​of the manufactured semiconductor device are associated with the substrate ID. Hereafter, the reference substrate... The characteristic values ​​of the semiconductor device fabricated on plate 21R are denoted as characteristic value 61R. The characteristic values ​​of a semiconductor device fabricated on a substrate having a specified process list are simply collected. It may be referred to as a "value." Also, on the board associated with process list 35 and the board ID The characteristic value of the fabricated semiconductor device will be denoted as characteristic value 65. In other words, characteristic value 65 is... These can be described as characteristic values ​​collected by the S001.

[0060] The above test compares characteristic value 61R with process list 35, and if the similarity is the same or the similarity is the same... The characteristic values ​​of the semiconductor device manufactured according to the highest process list are used. This is not limited to the above, but also includes characteristic value 61R and multiple characteristic values ​​collected in step 001, It is permissible to proceed. Furthermore, the above tests are performed for each characteristic value.

[0061] For the above tests, it is advisable to use methods such as the t-test.

[0062] If the above test results indicate that there is no significant difference between characteristic value 61R and characteristic value 65, Then proceed to step S003. Note that if there is a significant difference between characteristic value 61R and characteristic value 65 If this is determined, there is a possibility that the first lot 11 was not performed correctly. Therefore, it is not significant. If a significant difference is detected, verify whether the first lot 11 was executed correctly. We will notify the user. After notifying the user, we will terminate the process.

[0063] <<Step S003>> In step S003, an analysis of variance was performed on characteristic value 61, and the changes within the first lot 11 were made. Determine whether the process affects the characteristic value and record whether or not there is an effect. First, the first step Analysis of variance (Analysis of v) of characteristic values ​​61 between substrates 21 contained in the set 11. This is also called ariance or ANOVA. A comparison will be made between this and process list 31. Analysis of variance (ANOVA) is performed for each characteristic value.

[0064] The above analysis of variance includes Type 2 ANOVA, Type 3 ANOVA, and nonparametric analysis. Tests such as the Kruskal-Wallis test or the Friedman test are used.

[0065] Note that statistical analysis may be performed before conducting the above analysis of variance. By performing statistical analysis, Appropriate selection of methods for dispersion analysis is possible. For example, outlier detection can be performed as part of statistical analysis. This involves performing tests such as the normality of the distribution and the homogeneity of the variances. Through this statistical analysis, characteristics are identified. If the value 61 is determined to have no outliers, to be normal in its distribution, and to have homogeneous variances, For the above analysis of variance, it is advisable to choose a parametric test. Therefore, it is advisable to use Type 2 ANOVA. This improves the accuracy of the analysis of variance. It is possible.

[0066] Furthermore, the above statistical analysis revealed that characteristic value 61 has outliers, the distribution is not normal, and Alternatively, if it is determined that there is no homogeneity of variances, the above analysis of variance is nonparametric. It is advisable to select a hypothesis test. This will improve the accuracy of the analysis of variance. If there is no correspondence between characteristic values ​​61 between substrates 21, a non-parametric test is performed using Classca. We will use the Le-Wallis test.

[0067] For outlier detection, methods such as the Local Outlier Factor are used. For testing the authenticity of fabrics, there are the Shapiro-Wilk test, the Kolmogorov-Smirnov test, and others. It is used. In particular, the Shapiro-Wilk test is used. Also, for testing the homogeneity of variances, ruby Rubeen's test, Bartlett's test, Hartley's test, etc. are used. In particular, Rubeen's test is used. ru.

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

[0069] Furthermore, among the substrates 21 (substrates 21_1 to substrates 21_n) included in the first lot 11, By comparing process list 31 (process list 31_1 to process list 31_n), Different processes can be extracted between the plates 21.

[0070] Furthermore, in comparing the process lists, as explained in <<Step S001>>, the process list The processes and processing conditions (parameters) set in the table will be the subject of comparison. First, each process Check for match / mismatch. If the processes match, check one of the processing conditions (parameters) for that process. Check for matches / mismatches. Check for matches / mismatches in the processes and the processing conditions (parameters) of the processes. To check for matches or mismatches, you can use algorithms like the diff algorithm.

[0071] According to the above analysis of variance, the substrates 21 (substrates 21_1 to substrates) included in the first lot 11 Between 21_n, one or more characteristic values ​​of characteristic value 61 (characteristic value 61_1 to characteristic value 61_n) If a significant difference is determined, the different processes between the substrates 21 will have a significant impact on the characteristic value. It is recorded when an effect is applied. From here on, processes that are recorded as having a significant effect on characteristic values ​​will be treated as simple. This is sometimes referred to as a process that is affected. On the other hand, there is no significant difference in characteristic values ​​61 between substrates 21. If this is determined, it is recorded that the different processes between the substrates 21 do not significantly affect the characteristic values. do.

[0072] Furthermore, among the substrates 21 (substrates 21_1 to substrates 21_n) included in the first lot 11, Among the 61 characteristic values ​​(characteristic value 61_1 to characteristic value 61_n), those that are determined to have a statistically significant difference. The characteristic values ​​may include both characteristic values ​​that are deemed to have no significant difference. Therefore, the substrate Whether or not there is an impact on the characteristic values ​​of different processes between 21 may be recorded for each characteristic value.

[0073] After the above recording is complete, proceed to step S004.

[0074] <<Step S004>> In step S004, the substrate 21 (substrate 21_1 to substrate) included in the first lot 11 is Process list 31 (process list 31_1 to process list 31_n) prepared in 21_n) For all of them, a comparison is made with the process list group 30, and for each of the process list 31 Then, collect a list of processes from the group of 30 process lists that differ only in the processes that are affected. The process list collected in step S004 is referred to as process list 37. Process List 37 is a process list in which only the affected processes differ for each of the process lists 31. Also, in other words, step S004 can be rephrased as the process of obtaining process list 37. It is possible.

[0075] Furthermore, in comparing the process lists, as explained in <<Step S001>>, the process list The processes and processing conditions (parameters) set in the table will be the subject of comparison. First, each process Check for match / mismatch. If the processes match, check one of the processing conditions (parameters) for that process. Check for matches / mismatches. Check for matches / mismatches in the processes and the processing conditions (parameters) of the processes. To check for matches or mismatches, you can use algorithms like the diff algorithm.

[0076] For each of the process list 31, the process list differs only in the affected process, If one or more items are collected from the list group 30, proceed to step S005. From here on, The characteristic values ​​of semiconductor devices fabricated on a substrate associated with the process list 37 and the substrate ID are specified. This is denoted as a characteristic value of 67. In other words, characteristic value 67 is the characteristic value collected in step S004. It is possible.

[0077] <<Step S005>> In step S005, based on the parameters for the affected process and characteristic value 67, Machine learning is performed. For this machine learning, it is preferable to use, for example, regression analysis. Then, step S005 is a regression between the parameters for the affected process and the characteristic value 67. This can be described as the process of conducting analysis. By using regression analysis, the parameters of the process that has an influence can be analyzed. Therefore, the correlation between the characteristic value 67 and the semiconductor element can be analyzed. It can be estimated.

[0078] In step S005, specifically, the parameters for the affected process are used as explanatory variables. Then, regression analysis is performed with characteristic value 67 as the dependent variable. For example, the process with an influence... Linear regression is performed on the meter and characteristic value 67 using the least squares method.

[0079] Before performing the regression analysis described above, an analysis of variance (ANOVA) of characteristic value 67 was performed for the processes that were found to have an influence. You may also confirm that the difference in the parameters is significant with respect to the characteristic value. By confirming this, the accuracy of the semiconductor element characteristic values ​​estimated based on the results of the regression analysis above is It can be judged as expensive.

[0080] The above analysis of variance includes Type 2 ANOVA, Type 3 ANOVA, and nonparametric analysis. Tests such as the Kruskal-Wallis test or the Friedman test are used.

[0081] Note that statistical analysis may be performed before conducting the above analysis of variance. By performing statistical analysis, Appropriate selection of methods for dispersion analysis is possible. For example, outlier detection can be performed as part of statistical analysis. This involves performing tests such as the normality of the distribution and the homogeneity of the variances. Through this statistical analysis, characteristics are identified. If the value 67 is determined to have no outliers, normal distribution, and homogeneous variance, For the above analysis of variance, it is advisable to choose a parametric test. Therefore, it is advisable to use Type 2 ANOVA. This improves the accuracy of the analysis of variance. It is possible.

[0082] Furthermore, the above statistical analysis revealed that characteristic value 67 has outliers, and the distribution is not normal, Alternatively, if it is determined that there is no homogeneity of variances, the above analysis of variance is nonparametric. It is advisable to select a hypothesis test. This will improve the accuracy of the analysis of variance. If there is no correspondence between the characteristic value 67 and the boards associated with the process list 37 and the board ID, then The Kruskal-Wallis test is used as the parametric test.

[0083] For outlier detection, methods such as the Local Outlier Factor are used. For testing the authenticity of fabrics, there are the Shapiro-Wilk test, the Kolmogorov-Smirnov test, and others. It is used. In particular, the Shapiro-Wilk test is used. Also, for testing the homogeneity of variances, ruby Rubeen's test, Bartlett's test, Hartley's test, etc. are used. In particular, Rubeen's test is used. ru.

[0084] By performing the above analysis of variance, the differences in parameters for the affected processes are found to be related to the characteristic values. If the result is significant, then a regression analysis should be performed between the parameter and the characteristic value. By performing this analysis, the characteristic values ​​of the semiconductor device can be estimated from the relevant parameters.

[0085] Regression analysis includes linear regression, ridge regression, Lasso regression, elastic net, and k-negative regression. Parametric regression, regression trees, random forests, support vector regression, neural networks, etc. Which one should I use?

[0086] Note that the above example illustrates a method of performing analysis of variance before performing regression analysis, but alternatives to analysis of variance are... Furthermore, calculating the correlation coefficient using regression analysis can also serve as a criterion for deciding which semiconductor device estimates to adopt. The correlation coefficients are Pearson's product-moment correlation coefficient, Spearman's rank correlation coefficient, and Kendall's forward-phase correlation coefficient. You can use relational numbers or similar methods.

[0087] Based on the above, the characteristic values ​​of the semiconductor device can be obtained from the affected process and the parameters for that process. It can be estimated.

[0088] After step S005 is completed, the process list 31 and the characteristic value 61 will be These are stored in the process list group 30 and the characteristic value group 60. This storage is performed in step S This may be done after 006 is completed.

[0089] The above is an explanation of step S005.

[0090] Next, as an action to be performed by the user before step S006, the user shall perform the second step We will plan lot 12. Note that the second lot 12 will be planned after step S005 is completed. This is not limited to cases where the user is shown. The user is shown the second lot 12 when step S001 is opened. You can plan it before it starts, or after steps S001 to S005 have been implemented. You can start planning while you're waiting.

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

[0092] <<Step S006>> In step S006, based on the results of the regression analysis performed in step S005, the process list is created. From 32, we estimate the characteristic value 62. Then, we output the estimated characteristic value 62. In other words, Step S006 is the process of estimating the characteristic value 62 and outputting the estimated characteristic value 62. Here, process list 32 is the process list prepared for the second lot 12. Also, Characteristic value 62 is the characteristic value of the semiconductor device that is to be manufactured according to process list 32.

[0093] Note that the output is not limited to the estimated characteristic value 62. For example, the second lot 12 When only the processes that do not affect each other differ among the process lists 32 prepared for the substrate 22 included in the substrate 22. The system notifies the user that there is a risk that the difference in characteristic values ​​62 between the substrates 22 may become small. It is also possible to use a process lithography provided on the reference substrate 22R specified in the second lot 12. In step 32R, information regarding which process affects characteristic value 62 will be provided to the user. You may know.

[0094] After outputting the estimated characteristic value 62, or after notifying the user of the above information, the process will terminate. ru.

[0095] By performing steps S001 to S006, it is possible to avoid fabricating semiconductor devices. Alternatively, the characteristic values ​​of a semiconductor device can be estimated without measuring its electrical properties. This allows for a reduction in the number of times semiconductor devices need to be prototyped, thereby lowering development costs and shortening the development period. This allows for shortening the time between steps, etc.

[0096] Furthermore, by using the method for estimating the characteristic values ​​of semiconductor devices described in this section, the characteristic values ​​and processes can be calculated. Data such as process lists and process parameters are appropriately linked and stored. As data accumulates, the ability to estimate characteristic values ​​improves, resulting in more accurate characteristic values. This allows us to make an estimate. Also, by planning the second lot 12, process list 32 Since the process list group 30 is automatically compared, users can get an overview and analyze the data. It is possible.

[0097] The above is an explanation of one example of a method for estimating the characteristic values ​​of semiconductor devices.

[0098] <Example of a flow> This section provides an example of steps S001 through S006 as explained in the previous <flowchart>. This will be explained using Figures 3 through 8.

[0099] Figure 3 illustrates an example of step S001.

[0100] In step S001, as described above, the process list 31R and the process list group 30, A comparison will be made. For example, in Figure 3, among process list 30_1 to process list 30_m, If process list 30_2 is a process list whose similarity to process list 31R is above a certain level, At this time, process list 30_2 is collected. Note that process list 30_2 is shown in Figure 3. This corresponds to process list 35 explained in the previous <flow>. Then proceed to step S002. move on.

[0101] Figure 4 illustrates an example of step S002.

[0102] In step S002, as described above, a test is performed using characteristic value 61R and characteristic value 65. cormorant.

[0103] If a board ID has been assigned, for example, in Figure 4, the data collected in step S001 Based on the substrate ID associated with process list 30_2, select the characteristic value from the characteristic value group 60. 60_2 is extracted. Note that the characteristic value 60_2 shown in Figure 4 is explained in the <flow> above. This corresponds to characteristic value 65. And characteristic value 61R (characteristic value 61R(1) or characteristic value 61R( Test p)) against characteristic value 60_2 (characteristic value 60_2(1) to characteristic value 60_2(p)). The test is performed. The test determines that there is no significant difference between characteristic value 61R and characteristic value 60_2. If this occurs, proceed to step S003.

[0104] Figure 5 illustrates an example of step S003.

[0105] In step S003, as described above, analysis of variance of characteristic values ​​61 is performed between the substrates 21, and A comparison of the process list 31 is performed between the substrates 21. For example, in Figure 5, the process list 31 is shown between the substrates 21. By comparing process lists 31_1 to 31_n), it was found that there are different processes between the substrates 21. Therefore, it is determined to be process A. Note that in Figure 5, different processes between substrates 21 are considered to be process A. While this example illustrates one case, the different processes between substrates 21 may be two or more processes.

[0106] Furthermore, for example, in Figure 5, characteristic values ​​61 (characteristic value 61_1 to characteristic value 6) are shown between the substrates 21. Analysis of variance (1_n) indicates that there is a significant difference in the characteristic values ​​included in characteristic value 61. At this point, after process A is recorded as an affected process, the process proceeds to step S004. .

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

[0108] In step S004, as described above, for all of process list 31, process list A comparison was made with group 30, and for each of the process list 31, only the processes that were affected differed. A list of processes is collected from the group of process lists 30. For example, in Figure 6, the comparison is Therefore, process list 30_1 is a process list that differs from process list 31_1 only in process A. These are collected. Here, process list 30_1, etc., are the processes explained in the <flow> above. This corresponds to List 37. Note that each of process lists 31_2 through 31_n is... Similarly, assume that a list of processes is collected in which only process A is different. Then, step Proceed to S005.

[0109] Figure 7 illustrates an example of step S005.

[0110] In step S005, as described above, the parameters and characteristics of the affected process are determined. A regression analysis is performed with value 67. In Figure 7, characteristic value 67 is the amount collected in step S004. Characteristics of semiconductor devices fabricated on substrates associated with the substrate ID, such as in the program list 30_1. It is a value. In other words, characteristic value 67 is characteristic value 60_1, etc. Also, for example, in Figure 7 Analysis of variance revealed that in process list 31_1, the difference in parameters compared to process A was, It has been confirmed to be significant with respect to the characteristic value. At this time, the parameter for process A Then, regression analysis is performed on characteristic values ​​such as characteristic value 61_1 and characteristic value 60_1. After that, Proceed to S006.

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

[0112] As described above, in step S006, the results of the regression analysis performed in step S005 are... Based on this, characteristic value 62 is estimated from process list 32. For example, in Figure 8, step S00 Based on the results of the regression analysis performed in step 5, process list 32 (process list 32_1 to process list From (32_n), we estimate characteristic value 62 (characteristic value 62_1 to characteristic value 62_n). Then, after outputting the estimated characteristic value of 62, it terminates.

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

[0114] <Example of system configuration> This section describes an example configuration of a system for estimating the characteristic values ​​of a semiconductor device, which is one aspect of the present invention. This will be explained using Figure 9.

[0115] Figure 9 shows an example configuration of system 100 that can estimate the characteristic values ​​of semiconductor devices. The system 100 consists of an input unit (not shown in Figure 9), a database 110, and It has a processing unit 120. Note that the database 110 and the processing unit 120 are connected via a network. It is connected via a local area network. This includes LANs and the internet. Furthermore, the network can be wired, and Wireless communication, either one or both, can be used.

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

[0117] Database 110 contains process list group 30 prepared for substrate group 20, and substrate group 2 The set of 60 characteristic values ​​associated with the value 0 is stored here.

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

[0119] The processing unit 120A receives the process list 31 and characteristic values ​​61 via the input unit. Furthermore, the processing unit 120A receives the process list group 30 stored in the database 110. And characteristic value group 60 is input.

[0120] The processing unit 120A has the function to process steps S001 to S005 described above. It has the function of comparing two process lists and two or more. A function that performs calibration using the characteristic values ​​of semiconductor devices, parameters for the process, and two or more It has the function of performing regression analysis with the characteristic values ​​of semiconductor devices. Note that this test includes variance analysis. Analysis is included. Furthermore, the processing unit 120A may have a function for performing statistical analysis.

[0121] The two process lists mentioned above are the process list 31 and the database entered via the input unit. A selection is made from the process list group 30 stored in -110. Also, two or more of the above The characteristic value of the semiconductor device is the characteristic value 61 of the semiconductor device input via the input unit, and data The characteristic values ​​of the semiconductor elements are selected from the group of 60 stored in the base 110. The characteristic values ​​of two or more semiconductor devices used in the determination, and the two or more semiconductors used in the regression analysis. The characteristic values ​​of the physical element may not be identical.

[0122] Furthermore, the processing unit 120A may also have a function to output out1. Here, out1 This is information to notify the user that the first batch was executed correctly. Yes. By outputting this information, users can compare the first lot with past lots. Even without it, you can know that the first batch may not have been executed correctly.

[0123] The process list 32 is input to the processing unit 120B via the input unit. The results of the regression analysis performed in processing unit 120A are input to 20B.

[0124] The processing unit 120B has the function of processing step S006 described above. Unit 120B estimates characteristic values ​​62 from the process list 32 input via the input unit. To have the ability.

[0125] Furthermore, the processing unit 120B has a function to output out2. Here, out2 is a push This is the defined characteristic value 62, or the information described in <<Step S006>>.

[0126] The above configuration provides a system 100 capable of estimating the characteristic values ​​of semiconductor devices. It is possible to do so. Also, by using system 100, characteristic values, process list, and process Data such as the corresponding parameters are appropriately linked and stored. Therefore, the data is stored. As this process continues, the ability to estimate characteristic values ​​improves, leading to more accurate estimation of characteristic values. This can be done. Also, by planning the second lot 12, process list 32 and process list group 3 Because it is automatically compared to 0, users can get an overview and analyze the data.

[0127] <Details of the system configuration example> This section describes in detail an example configuration of system 100, which is one aspect of the present invention, using Figure 10. To clarify. In the following, unless otherwise specified, the components of the system described in this section, The functions of the component are the components of system 100 and the functions of the component. These and other descriptions can be taken into consideration.

[0128] Figure 10 shows a system 100A that can estimate the characteristic values ​​of semiconductor devices. System 100A is a detailed description of the configuration of System 100 shown in Figure 9. System 100A, like System 100, has an input section (not shown in Figure 10) and a database It has a -110 and a processing unit 120. The database 110 is a storage unit 111 , and storage unit 112. The processing unit 120 also has processing units 121 to 124 It holds.

[0129] <<Storage section 111>> The memory unit 111 stores the process list group 30. The memory unit 111 also stores the process list group 30. The pre-prepared parameter set is stored. Note that the memory unit 111 contains <<step As explained in S003>>, the different processes between substrates 21 and whether or not they affect the characteristic value 61 It may be recorded.

[0130] <<Storage section 112>> The memory unit 112 stores the characteristic value group 60.

[0131] <<Processing step 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 to compare process lists. The comparison of process lists is as described above. As shown above, it is good to use the diff algorithm. Therefore, the diff algorithm is It is preferable that the data be stored in the memory unit (not shown in Figure 10) of the processing unit 121. Alternatively, If the diff algorithm is stored in the memory unit 111, then the diff algorithm However, it is preferable that the data be supplied to the processing unit 121 from the memory unit 111 or the like.

[0133] For example, the processing unit 121 is the process list 31R described in <<Step S001>> and The process list group 30 and can be compared. The processing unit 121 compares the process list 31R and A list of processes with a similarity of a certain level or higher is collected from the group of process lists 30. Then, processing Section 121 contains the collected process list, or the substrate associated with the collected process list. The ID is output to the processing unit 122. The process list output to the processing unit 122 is <<Step This corresponds to process list 35 explained in S001>>.

[0134] Furthermore, for example, the processing unit 121 is the first lot as described in <<Step S003>> The process list 31 can be compared between the substrates 21 included in 11. Processing unit 121 The processing unit 121 extracts different processes between the substrates 21. The process, or the process ID of that process, is output to the processing unit 122.

[0135] Furthermore, for example, the processing unit 121 is the process listed in step 3, as explained in step S004. A comparison can be made between 1 and the process list group 30. The processing unit 121 can perform a comparison with process list 3 For each of the 1s, a process list is created in a group of 30 process lists, where only the affected process differs. Then, the processing unit 121 collects the collected process list, or the collected process list. The board ID associated with the strip is output to the processing unit 123. The process list corresponds to process list 37 described in <<Step S004>>.

[0136] <<Processing step 122>> The characteristic value 61 is input to the processing unit 122. Also, the output from the processing unit 121 is The process list, or the substrate ID associated with the process list, is entered. The output from 121 indicates different processes between the boards 21, or the process ID of the said process is input. ru.

[0137] The processing unit 122 assigns characteristic values ​​corresponding to the above process list or the above substrate ID to characteristic value group 6 It has a function to collect data starting from 0. Furthermore, when the characteristic value group 60 is input to the processing unit 122, the above process Characteristic values ​​corresponding to the sequence list or the above-mentioned substrate ID may be extracted from the characteristic value group 60.

[0138] Furthermore, the processing unit 122 has a function to perform characteristic value testing. This testing includes variance testing. Analysis is included. Furthermore, the processing unit 122 has the function of outputting out1. Part 122 may have a function for performing statistical analysis. Note that it may also have functions for analysis of variance, statistical analysis, etc. The golism is preferably stored in the memory unit of the processing unit 122 (not shown in Figure 10). Alternatively, if algorithms such as analysis of variance or statistical analysis are stored in the memory unit 112 or the like. Algorithms such as analysis of variance and statistical analysis are supplied from the memory unit 112 to the processing unit 122. It would be good to do so.

[0139] For example, the processing unit 122 uses the characteristic value 61R, as explained in <<Step S002>>, and The characteristic values ​​collected by TEP S001 can be used for testing. Based on the results of this test, If a discrepancy is detected, the processing unit 122 outputs out1. The characteristic value collected in S001 corresponds to characteristic value 65 explained in <<Step S002>>. In addition, out1 is the information to be notified to the user, as explained in <<Step S002>>. It is as follows.

[0140] Also, when it is determined that there is no significant difference as a result of the above inspection, the processing unit 122 can perform the analysis of variance as described in <<Step S003>>. At this time, it has a function of outputting to the storage unit 111 whether there is an influence on the characteristic values of different processes between the substrates 21. Note that the processing unit 122 may perform statistical analysis as described in <<Step S003>>. between substrates 21 for the presence or absence of influence on the characteristic values of different processes, and outputting to the storage unit 111. Note that the processing unit 122 may perform statistical analysis as described in <<Step S003>>. It has a function of outputting to the storage unit 111 whether there is an influence on the characteristic values of different processes between the substrates 21. Note that the processing unit 122 may perform statistical analysis as described in <<Step S003>>. The processing unit 122 may perform statistical analysis as described in <<Step S003>>.

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

[0142] The processing unit 123 has a function of collecting the characteristic values corresponding to the above process list or the above substrate 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 above process list or the above substrate ID may be extracted from the characteristic value group 60. The processing unit 123 has a function of collecting the characteristic values corresponding to the above process list or the above substrate 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 above process list or the above substrate ID may be extracted from the characteristic value group 60. The characteristic value group 60 may be input to the processing unit 123, and the characteristic values corresponding to the above process list or the above substrate 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 parameter and the characteristic value for the process with an influence as described in <<Step S005>>. Also, the processing unit 123 has a function of outputting the result of the regression analysis to the processing unit 124. Also, the processing unit 123 has a function of outputting the process list 31 to the storage unit 111. Also, the processing unit 123 has a function of outputting the characteristic value 61 to the storage unit 112. Note that the regression analysis algorithm may be stored in a storage unit (not shown in FIG. 10) that the processing unit 123 has. Or, when the regression analysis algorithm is stored in the storage unit 112 or the like, the regression analysis algorithm is stored in the storage unit 11 The processing unit 123 has a function of performing a regression analysis between the parameter and the characteristic value for the process with an influence as described in <<Step S005>>. Also, the processing unit 123 has a function of outputting the result of the regression analysis to the processing unit 124. Also, the processing unit 123 has a function of outputting the process list 31 to the storage unit 111. Also, the processing unit 123 has a function of outputting the characteristic value 61 to the storage unit 112. Note that the regression analysis algorithm may be stored in a storage unit (not shown in FIG. 10) that the processing unit 123 has. Or, when the regression analysis algorithm is stored in the storage unit 112 or the like, the regression analysis algorithm is stored in the storage unit 11 The processing unit 123 has a function of performing a regression analysis between the parameter and the characteristic value for the process with an influence as described in <<Step S005>>. Also, the processing unit 123 has a function of outputting the result of the regression analysis to the processing unit 124. Also, the processing unit 123 has a function of outputting the process list to the storage unit 111. Also, the processing unit 123 has a function of outputting the characteristic value 61 to the storage unit 112. Note that the regression analysis algorithm may be stored in a storage unit (not shown in FIG. 10) that the processing unit 123 has. Or, when the regression analysis algorithm is stored in the storage unit 112 or the like, the regression analysis algorithm is stored in the storage unit 11 The processing unit 123 has a function of outputting the process list 31 to the storage unit 111. Also, the processing unit 123 has a function of outputting the characteristic value 61 to the storage unit 112. Note that the regression analysis algorithm may be stored in a storage unit (not shown in FIG. 10) that the processing unit 123 has. Or, when the regression analysis algorithm is stored in the storage unit 112 or the like, the regression analysis algorithm is stored in the storage unit 11 The processing unit 123 has a function of outputting the characteristic value 61 to the storage unit 112. Note that the regression analysis algorithm may be stored in a storage unit (not shown in FIG. 10) that the processing unit 123 has. Or, when the regression analysis algorithm is stored in the storage unit 112 or the like, the regression analysis algorithm is stored in the storage unit 11 The regression analysis algorithm may be stored in a storage unit (not shown in FIG. 10) that the processing unit 123 has. Or, when the regression analysis algorithm is stored in the storage unit 112 or the like, the regression analysis algorithm is stored in the storage unit 11 The regression analysis algorithm may be stored in a storage unit (not shown in FIG. 10) that the processing unit 123 has. Or, when the regression analysis algorithm is stored in the storage unit 112 or the like, the regression analysis algorithm is stored in the storage unit 11 It is preferable that the power is supplied to the processing unit 123 from source 2, etc.

[0144] The processing unit 123 may also have a function for performing characteristic value testing. This testing includes variance testing. Analysis is included. Furthermore, the processing unit 123 may have a function for performing statistical analysis. Processing unit 1 If 23 does not have the function to perform characteristic value testing and statistical analysis, <<Ste The analysis of variance and statistical analysis described in section S005>> are performed in processing unit 122. Note that in Figure 10, data, instructions, etc. are input between processing unit 122 and processing unit 123. The arrows corresponding to the output are not shown in the diagram.

[0145] <<Processing section 124>> Process list 32 is input to processing unit 124. Also, output from processing unit 123 The results of the regression analysis are then entered.

[0146] The processing unit 124 uses the results of the regression analysis described above to estimate characteristic values ​​from the process list 32. It has the function of outputting out2. The characteristic value estimated from St32 is the characteristic value 62 explained in <<Step S006>>. It corresponds to the estimated characteristic value and the explanation given in <<Step S006>>. This includes information to be notified to users, etc.

[0147] Note that until process list 32 is entered into processing unit 124, the process performed in processing unit 123 will be recorded. The results of the regression analysis are stored in the temporary storage area of ​​the processing unit 123 or processing unit 124. It would be good to do this. Alternatively, the results of the regression analysis performed in the processing unit 123 can be stored in the memory unit 111, etc. Stored, and when process list 32 is input to processing unit 124, processing unit 124 It can be entered here.

[0148] With this configuration, it is possible to provide a system that can estimate the characteristic values of semiconductor elements. It is possible.

[0149] <Modified Example of System> Note that the configuration example of the system for estimating the characteristic values of semiconductor elements is not limited to the configuration of the system 100A shown in FIG. 10. Hereinafter, a modified example of the system for estimating the characteristic values of semiconductor elements will be described using FIGS. 11 and 12. values will be described.

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

[0151]

[0152] The system 100B shown in FIG. 11 is a modified example of the system 100A shown in FIG. 10. The system 100B is different from the system 100A in that it has a storage unit 113.

[0153] The storage unit 113 stores the results of the regression analysis performed by the processing unit 123. As a result, the system 100B can be put into a standby state until the process list 32 is input.

[0154] <<Modified Example 2 of System>> FIG. 12 shows a configuration example of the system 100C. In the system 100C shown in FIG. 12, elements having the same functions as the elements constituting the system shown in <Configuration Example of System> and <Details of Configuration Example of System> are denoted by the same reference numerals.

[0155] System 100C shown in Figure 12 is a modified version of System 100A shown in Figure 10. Stem 100C is the same as System 100A as described above, in steps S001 to S0 Before performing 06, the process list 31 is stored in the memory unit 111, and the characteristic value 61 is stored in the memory unit 1 The difference is that it is stored in 12.

[0155] With the above configuration, the process list and characteristic values ​​of the implemented lot are sequentially stored in the database. By being stored in the system, before planning the second lot 12, the planning of the first lot 11 and System 100C can be used without implementing the above. The tasks that the user is supposed to perform before step S001 do not need to be performed.

[0156] For example, the process list of the reference substrate is selected from the process list group 30. Next, via the processing unit 121, processes with a similarity of above a certain level are selected for the process list. The process list is collected from the process list group 30. The collected process list is used for the first lot 1. The process list prepared in step 1 is designated as process list 31R, and the selected process list is designated as process list 31R. It is recommended to use Stem 100C. This will allow for the planning and implementation of the first lot 11. Even without it, you can still use System 100C.

[0157] <Computer equipment> This section describes a system for estimating the characteristic values ​​of a semiconductor device, which is one aspect of the present invention. The computer device will be explained using Figure 13.

[0158] Figure 13 illustrates a computer device having a system for estimating the characteristic values ​​of semiconductor elements. This is a diagram. Computer device 1000 accesses database 1011 via the network. It is connected to remote computer 1012 and remote computer 1013. The computer device 1000 includes an arithmetic unit 1001, a memory 1002, and an input / output interface. The computer has a memory module 1003, a communication device 1004, and a storage device 1005. The device 1000 communicates with the display device 1006a via the input / output interface 1003. It is electrically connected to the keyboard 1006b. Furthermore, the computer device 1000 is The communication device 1004 is electrically connected to the network interface 1007. The network interface 1007 is connected via the network. Database 1011, remote computer 1012, and remote computer 1 It is electrically connected to 013.

[0159] Here, the above network includes local area networks (LANs) and internet networks. This includes the internet. Furthermore, the above network can be either wired or wireless. It can use communication by both methods. In addition, wireless communication can be used in the above network. If present, use short-range communication such as Wi-Fi (registered trademark) or Bluetooth (registered trademark). In addition to the means, there are communication methods compliant with the third-generation mobile communication system (3G), and LTE (3.9G). A communication method compliant with (sometimes called) the fourth-generation mobile communication system (4G) Various communication methods, such as communication methods compliant with the fifth-generation mobile communication system (5G). They can use means of communication.

[0160] In a system for estimating characteristic values ​​of a semiconductor device, according to one aspect of the present invention, a database 110 corresponds to database 1011. Note that database 110 is storage It may also be 1005. Also, process list 31 and characteristic value 61 are initially stored Stored in page 1005, and after step S005 described above is completed, database 101 It may be stored in 1.

[0161] Furthermore, the processing unit 120 corresponds to the arithmetic unit 1001. The arithmetic unit of the remote computer 1012 or the remote computer 1013 It is also possible that the processing units 121 to 123 of the processing unit 120 are remote control A computing device of computer 1012, or remote computer 1013, and processing The processing unit 124 in unit 120 may be an arithmetic unit 1001.

[0162] Furthermore, the aforementioned out1 and out2 are displayed on the display device 1006a. Oh, the display device 1006a, the comparison results of the process list mentioned above, the test results, analysis of variance As a result, the results of statistical analysis, etc., are converted into and displayed as tables, formulas, graphs, etc. That's fine.

[0163] Based on the above, one aspect of the present invention provides a method for estimating the characteristic values ​​of a semiconductor device to be prototyped. It can be used. Furthermore, one aspect of the present invention predicts the characteristic values ​​of a semiconductor device to be prototyped. A system can be provided that defines the semiconductor. Furthermore, one aspect of the present invention is the semiconductor prototyped in this invention. We compare the process and characteristic values ​​of the element with the process and characteristic values ​​of semiconductor elements that were prototyped in the past. We can provide a system that performs operations dynamically.

[0164] This embodiment can be implemented by combining parts of it as appropriate. [Explanation of Symbols]

[0165] :11:First lot, 12:Second lot, 20:Group of boards, 20_m:Board, 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 T, 30_2: Process list, 31: Process list, 31_n: Process list, 31_1: Process List, 31_2: Process list, 31R: Process list, 32: Process list, 32_n: Process 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, 6 1: Characteristic value, 61_n: Characteristic value, 61_1: Characteristic value, 61_2: Characteristic value, 61R: Characteristic value , 62: characteristic value, 62_n: characteristic value, 62_1: characteristic value, 65: characteristic value, 67: characteristic value, 100: System, 100A: System, 100B: System, 100C: System, 1 10: Database, 111: Storage unit, 112: Storage unit, 113: Storage unit, 120: Processing unit Section, 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: S Treasury, 1006a: Display device, 1006b: Keyboard, 1007: Network Interface, 1011: Database, 1012: Remote computer, 1013 : Remote computer

Claims

1. It has an input unit, a database, and a processing unit. The input unit receives a first process list, a second process list, and characteristic values ​​of semiconductor devices. The database has the function of storing the first process list, the characteristic values ​​of the semiconductor device, the process list group, and the characteristic value group of the semiconductor device. The aforementioned processing unit, A function for comparing two process lists selected from the first process list and the group of process lists, A function to perform testing using the characteristic values ​​of two or more semiconductor devices selected from the characteristic values ​​of the semiconductor device and the group of characteristic values ​​of the semiconductor device, A function to perform regression analysis between parameters for a process and the characteristic values ​​of two or more semiconductor elements selected from the characteristic values ​​of the semiconductor elements and the group of characteristic values ​​of the semiconductor elements, A semiconductor element characteristic value estimation system having a function for estimating the characteristic values ​​of a semiconductor element from the second process list.

2. In claim 1, For the aforementioned comparison, we used the diff algorithm. Of the aforementioned tests, As a test using the characteristic values ​​of two semiconductor elements, we use the t-test. A semiconductor element characteristic value estimation system that uses a non-parametric test as a test using the characteristic values ​​of three or more semiconductor elements.

3. In claim 1 or claim 2, The database comprises a second storage unit and a third storage unit. The second storage unit has the function of storing the process list group, The third memory unit is a semiconductor element characteristic value estimation system having the function of storing a group of characteristic values ​​of the semiconductor element.

4. In any one of claims 1 to 3, The aforementioned processing unit, A first processing unit having the function of performing the aforementioned comparison, A second processing unit having the function of performing the aforementioned verification, A third processing unit having the function of performing the aforementioned regression analysis, A semiconductor element characteristic value estimation system comprising: a fourth processing unit having the function of estimating the characteristic value of a semiconductor element from the second process list.

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

Images