阅读说明：本技术 检查装置调整系统以及检查装置调整方法 (Inspection device adjustment system and inspection device adjustment method ) 是由 前田太一 笹气裕子 于 2019-05-15 设计创作，主要内容包括：本发明的课题在于,迅速确定检查对象即晶片的光学条件,特别在于,在获得客户晶片后,将光学条件设定迅速化。本发明所涉及的检查装置自动调整系统具有：输入解析条件的解析条件设定IF(102)；进行解析的解析执行部(103)；解析中所用的检查装置模型和模型DB(101)；蓄积解析结果的解析结果DB(104)；输入晶片图案、聚焦点、最优化指标和优先级的观察条件设定IF(105)；检索与所输入的晶片图案类似的晶片图案的晶片图案检索部(106)；从解析结果DB(104)提取类似晶片图案和在聚焦点中最优的光学条件的光学条件提取部(107)；和生成与光学条件对应的控制信号并发送到检查装置的光学条件设定部(108)。(The invention aims to quickly determine the optical condition of a wafer as an inspection object, and particularly to quickly set the optical condition after a customer wafer is obtained. The inspection device automatic adjustment system according to the present invention includes: an analysis condition setting IF (102) for inputting an analysis condition; an analysis execution unit (103) for performing analysis; an inspection device model and a model DB (101) used for analysis; an analysis result DB (104) for accumulating analysis results; inputting an observation condition setting IF (105) of a wafer pattern, a focus point, an optimization index and a priority; a wafer pattern search unit (106) for searching for a wafer pattern similar to the inputted wafer pattern; an optical condition extraction unit (107) for extracting a wafer-like pattern and an optical condition optimal for the focusing point from the analysis result DB (104); and an optical condition setting unit (108) for generating a control signal corresponding to the optical condition and transmitting the control signal to the inspection device.)

1. An inspection device adjustment system for adjusting an inspection device for inspecting a sample,

the inspection device adjustment system is characterized in that,

an analysis result database storing analysis results including, for each sample, an inspection apparatus model representing a model of an inspection apparatus, a sample model representing a model of a sample, a focus point representing a position at which the inspection apparatus focuses on the sample, a sample image of the sample, and an optical condition in the inspection apparatus of the sample,

the inspection device adjustment system includes:

a unit that receives an inspection sample image representing an image of a sample to be inspected from the inspection apparatus;

a unit that retrieves an analysis result having a sample image similar to the received inspection sample image from the analysis result database;

means for extracting optical conditions included in the retrieved analysis result; and

a unit that outputs the extracted optical condition.

2. The inspection device adjustment system of claim 1,

the unit that retrieves the analysis result retrieves a plurality of analysis results from the analysis result database,

the unit that extracts the optical condition determines an analysis result having similar sample images from the plurality of analysis results retrieved, and extracts the optical condition from the determined analysis result.

3. The inspection device adjustment system of claim 2,

further storing the performance under examination in said parsing result database,

the inspection apparatus adjustment system further includes:

a unit that accepts designation of an inspection apparatus model, a sample model, and a focus point; and

a unit that accepts designation of a performance in the prioritized inspection,

the unit that extracts the optical condition determines an analysis result having similar sample images from the plurality of analysis results following the received priority, extracts the optical condition from the determined analysis result.

4. The inspection device adjustment system of claim 3,

the unit retrieving the analysis result extracts the optical condition using multi-objective optimization or machine learning, following the priority.

5. The inspection apparatus adjustment system according to any one of claims 1 to 4,

the unit that extracts the optical condition performs:

setting a plurality of estimated values in which the deviation of the size of the sample to be inspected is estimated,

extracting a plurality of optical conditions corresponding to the plurality of set estimated values,

one of the extracted plurality of optical conditions is determined.

6. A method for adjusting a testing device, which uses a system for adjusting a testing device for performing a test on a sample,

the inspection apparatus adjustment system is connectable to an analysis result database storing analysis results including, for each sample, an inspection apparatus model representing a model of an inspection apparatus, a sample model representing a model of a sample, a focus point representing a position at which the inspection apparatus focuses on the sample, a sample image of the sample, and an optical condition in the inspection apparatus of the sample,

the inspection device adjustment method includes:

a step of receiving an inspection sample image representing an image of a sample to be inspected from the inspection apparatus;

a step of retrieving an analysis result having a sample image similar to the received inspection sample image from the analysis result database;

extracting an optical condition included in the retrieved analysis result; and

a step of outputting the extracted optical condition.

7. The inspection apparatus adjusting method according to claim 6,

the step of detecting the parsing result retrieves a plurality of parsing results from the parsing result database,

the step of extracting the optical condition determines an analysis result having similar sample images from the plurality of retrieved analysis results, and extracts the optical condition from the determined analysis result.

8. The inspection apparatus adjusting method according to claim 7,

further storing the performance under examination in said parsing result database,

the inspection apparatus adjustment method further includes:

receiving the designation of the inspection apparatus model, the sample model, and the focus point; and

a step of accepting designation of a performance in the prioritized inspection,

the step of extracting the optical condition determines an analysis result having similar sample images from the plurality of analysis results following the received priority, and extracts the optical condition from the determined analysis result.

9. The inspection apparatus adjusting method according to claim 8,

the step of detecting the analytical result follows the priority, and the optical condition is extracted using multi-objective optimization or machine learning.

10. The inspection apparatus adjusting method according to any one of claims 6 to 9,

the step of extracting the optical condition performs:

setting a plurality of estimated values in which the deviation of the size of the sample to be inspected is estimated,

extracting a plurality of optical conditions corresponding to the plurality of set estimated values,

one of the extracted plurality of optical conditions is determined.

11. An inspection apparatus for performing an inspection of a sample, the inspection apparatus being characterized in that,

an analysis result database storing analysis results including, for each sample, an inspection apparatus model representing a model of an inspection apparatus, a sample model representing a model of a sample, a focus point representing a position at which the inspection apparatus focuses on the sample, a sample image of the sample, and an optical condition in the inspection apparatus of the sample,

the inspection apparatus includes:

a unit that captures an inspection sample image representing an image of a sample to be inspected from the inspection apparatus;

a unit that retrieves an analysis result having a sample image similar to the imaged inspection sample image from the analysis result database;

means for extracting optical conditions included in the retrieved analysis result; and

a unit that outputs the extracted optical condition.

12. The inspection device of claim 11,

the unit that retrieves the analysis result retrieves a plurality of analysis results from the analysis result database,

the unit that extracts the optical condition determines an analysis result having similar sample images from the plurality of analysis results retrieved, and extracts the optical condition from the determined analysis result.

13. The inspection device of claim 12,

further storing the performance under examination in said parsing result database,

the inspection apparatus further includes:

a unit that accepts designation of an inspection apparatus model, a sample model, and a focus point; and

a unit that accepts designation of a performance in the prioritized inspection,

the unit that extracts the optical condition determines an analysis result having similar sample images from the plurality of analysis results following the received priority, extracts the optical condition from the determined analysis result.

14. The inspection device of claim 13,

the unit retrieving the analysis result extracts the optical condition using multi-objective optimization or machine learning, following the priority.

15. The inspection apparatus according to any one of claims 11 to 14,

the unit for extracting the optical condition performs the following operations:

setting a plurality of estimated values in which the deviation of the size of the sample to be inspected is estimated,

extracting a plurality of optical conditions corresponding to the plurality of set estimated values,

one of the extracted plurality of optical conditions is determined.

Technical Field

The present invention relates to a technique for automatically adjusting an inspection apparatus for a sample. In addition, the sample includes a semiconductor wafer (hereinafter also referred to as a wafer). In addition, the inspection apparatus further includes an observation apparatus.

Background

An examination apparatus for obtaining an image by irradiating a sample with an Electron beam (1 st Electron) and detecting a signal Electron (2 nd Electron) generated secondarily from the sample includes a Scanning Electron Microscope (SEM). In particular, an inspection apparatus for inspecting a semiconductor wafer uses a SEM, and one of them is a CD-SEM (Critical Dimension SEM). Miniaturization of semiconductors is advancing every year, and demands for CD-SEM, such as improvement of resolution and measurement reproducibility, and reduction of measurement errors between inspection apparatuses, are increasing. In order to obtain a high-resolution SEM image, it is necessary to accurately set optical conditions when the wafer is irradiated with an electron beam. The optical conditions include, for example, an acceleration voltage of an electron beam, an excitation current of a lens, and voltages of various electrodes. The optimum values of these conditions differ depending on the specifications of the wafer, the observation position, and the focus point. There are various kinds of wafers and patterns, and when there is a large difference in the kind of wafer to be inspected, it is necessary to adjust the optimum optical conditions again. Further, since the adjustment work is a complicated work for dealing with a plurality of optical conditions, it is difficult to perform the adjustment work except for a person who understands the mechanism of the target CD-SEM in detail. That is, even when the inspection wafer is largely changed after the CD-SEM is delivered to the customer, it is necessary for a professional to go to the customer to adjust the optical conditions. Since the CD-SEM is loaded into the production line, stopping the CD-SEM for adjustment reduces the wafer production efficiency. Therefore, automation and speediness of optical condition setting are required for the CD-SEM.

Patent document 1 discloses a method for automatically optimizing optical conditions using analysis and an observation image of a wafer. This is a method of obtaining an optimal optical condition in the inspection apparatus by comparing the analysis result with the observation image and correcting the analysis model. When the wafer surface has minute undulations, the height changes when the observation position changes, and therefore the optimal optical condition changes, but if there is a corrected analytical model, the optimal optical condition can be predicted with respect to the height.

Patent document 2 discloses a method of automatically searching for an observation position of a wafer using a characteristic image of the wafer captured in advance as a model. This is a method as follows: when searching for a specific observation position, a position matching a previously captured image is detected by pattern recognition, optical conditions at the time of capturing are stored, and the optical conditions are automatically set when the observation position is detected.

Documents of the prior art

Patent document

Patent document 1: JP 2006-32202A

Patent document 2: JP-A09-245709

Disclosure of Invention

Problems to be solved by the invention

In patent document 1, an analytical model is corrected using an observation image of a wafer, and an optimal optical condition is predicted. However, this method cannot be applied unless a wafer is obtained, because it requires an image to be observed.

Patent document 2 also requires an image of the wafer, and cannot be applied unless the wafer is obtained. In addition, since only limited optical conditions used for image capturing are stored, optimal optical conditions are not necessarily obtained at positions other than the image capturing position.

The present invention aims to predict the trend of semiconductor development and customer information, create an analysis model of a wafer, predict optical conditions by analysis, and give priority to and present the optical conditions as long as the specifications of the customer wafer can be grasped, thereby quickly setting the optical conditions.

Means for solving the problems

In order to solve the above problem, a database storing analysis results in which an inspection apparatus model representing a model of an inspection apparatus, a sample model representing a model of a sample, and a focusing point are associated with each other is used to determine optical conditions in inspection/imaging of the sample in the inspection apparatus. More specifically, the corresponding image data (pattern) is stored in the analysis result, an image similar to the image data (pattern) captured by the inspection apparatus is searched for, and the optical condition indicating the analysis result of the searched image is determined.

In addition, the present invention further comprises: candidates of optical conditions that comply with the designated priority order are extracted.

ADVANTAGEOUS EFFECTS OF INVENTION

According to the present invention, the optical conditions can be set more efficiently.

Drawings

Fig. 1 is a configuration diagram of an automatic adjustment system of an inspection apparatus according to embodiment 1.

FIG. 2 is a diagram showing an example of the structure of a CD-SEM.

Fig. 3 is a diagram showing a flow of use of the automatic adjustment system of the inspection apparatus according to embodiment 1.

Fig. 4 is an example of an input screen of the inspection apparatus automatic adjustment system according to embodiment 1.

Fig. 5 is an example of the analysis result DB104 of the inspection apparatus automatic adjustment system according to embodiment 1.

Fig. 6 is a diagram showing a flow of use of the automatic adjustment system of the inspection apparatus according to embodiment 1.

Fig. 7 is an example of an input screen of the inspection apparatus automatic adjustment system according to embodiment 1.

Fig. 8 is a diagram showing a flow of extracting the optimum optical condition in the inspection apparatus automatic adjustment system according to embodiment 1.

Fig. 9 is a configuration diagram of an automatic adjustment system of an inspection apparatus according to embodiment 2.

Fig. 10 is a diagram showing a flow for coping with dimensional variations according to example 3.

Detailed Description

After the customer wafer is obtained, in order to quickly set the optimum optical conditions, it is effective to create an analysis model of the wafer based on the prediction of the material and pattern of the wafer in advance and to obtain the optical conditions by analysis. Therefore, during or after the design of the inspection apparatus, analysis is performed before the customer wafer is obtained, and an analysis result DB is created. The analytical model includes a sample model relating to the sample and an inspection apparatus model for the inspection apparatus. Further, the sample model includes a wafer model for a semiconductor wafer. The term "optimum" in the present specification includes a concept conforming to a given condition (index), and is not necessarily a term having a meaning that is most appropriate.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are given to the common portions, and redundant description is omitted.

Example 1

Fig. 1 shows an automatic adjustment system of an inspection apparatus according to embodiment 1. The inspection apparatus automatic adjustment system includes a model DB101, an analysis condition setting IF102, an analysis execution unit 103, an analysis result DB104, an observation condition setting IF105, a wafer pattern search unit 106, an optical condition extraction unit 107, and an optical condition setting unit 108. The model DB101, analysis condition setting IF102, analysis execution unit 103, and analysis result DB104 are used as an analysis result DB creation unit 109, and the observation condition setting IF105, wafer pattern search unit 106, optical condition extraction unit 107, and optical condition setting unit 108 are used as an optimal optical condition extraction unit 110.

The structure of the analysis result DB creating unit 109 will be described. The analysis result DB creation unit 109 sets analysis conditions in the analysis condition setting IF102 using the analysis model of the inspection apparatus of the model DB101 and the analysis model of the wafer, executes the analysis in the analysis execution unit 103, and stores the results in the analysis result DB 104. The model DB101 is a DB in which analysis models of inspection apparatuses and analysis models of wafers used for analysis are created and accumulated in advance. The model of the inspection apparatus is an analytical model of the inspection apparatus designed by the manufacturer of the inspection apparatus and delivered to the customer. The analytical model of the inspection apparatus is an analytical model of an optical system or a mechanical system that is used by a designer at the time of design, for example, and is a simple model represented by an approximate expression or a large-scale model using a technique such as a finite element method. In the design stage of the inspection apparatus, it is difficult to completely grasp the specifications of the customer wafer. Therefore, a wafer model is created in a plurality of types based on the movement of semiconductor devices and customer information. The analysis condition setting IF102 is a screen to which an optical condition, a focus point, an inspection apparatus model that is a model of an inspection apparatus used for analysis, and a wafer model that is a model of a wafer are input as analysis conditions. The details will be described later. The analysis execution unit 103 extracts the inspection apparatus model and the wafer model input to the analysis condition setting IF from the model DB101, and executes analysis using the optical conditions and the focus point as inputs. The analysis result DB104 stores the analysis result obtained from the analysis execution unit 103 in association with the analysis conditions and the analysis model.

The configuration of the optimum optical condition extraction unit 110 will be described. The observation condition setting IF105 is a screen to which the specification of the wafer to be observed, the focal point as the observation position, the index of the performance optimized in observation, and the priority of each are inputted. The details will be described later. The wafer pattern search unit 106 searches the analysis result DB104 for a wafer pattern of a wafer similar to the wafer pattern to be observed. The optical condition extraction unit 107 extracts the retrieved similar wafer patterns and the optimum optical conditions corresponding to the focusing points. The extraction process of the optimal optical conditions will be described in detail later. The optical condition setting unit 108 outputs the optical condition extracted by the optical condition extraction unit 107. The output further comprises: generating a device control signal corresponding to the optical condition and setting the inspection device; the optical condition is displayed on a display screen. In the latter case, the optical condition is set to the inspection apparatus by the input of the user referring to the display content of the display screen. The optimum optical condition extraction unit 110 is provided in the inspection apparatus or connected thereto, although not shown. The analysis result DB creating unit 109 and the optimum optical condition extracting unit 110 are realized in an information processing device, and each function is executed in the processing device like a CPU that performs calculation in accordance with a program.

As described above, in the present embodiment, the customer wafer pattern is predicted before the customer wafer is obtained, an analysis model is created, and the optimal optical condition is predicted by analysis. This enables the optimum optical conditions to be found quickly. For example, the optical conditions can be found as long as the customer wafer pattern is clear.

The following will specifically illustrate a processing method of the inspection apparatus automatic adjustment system according to embodiment 1, with reference to a CD-SEM shown in fig. 2 as an inspection apparatus. FIG. 2 is a cross-sectional view showing an example of the structure of a CD-SEM. The CD-SEM includes a lens barrel 201 with a vacuum maintained therein. The lens barrel 201 includes: an electron source 202; a 1 st condenser lens 203 for condensing the electron beam; a 2 nd condenser lens 204; a deflector 205 for deflecting the electron beam; an objective lens 206 for converging the electron beam on the sample 207; a stage 208 on which the sample is held; and a detector 209 for detecting electrons. The 1 st-order electrons 210 irradiated from the electron source 202 are converged on the image planes C1, C2 by the objective lens while controlling the current value and the like, and are converged on the sample by the objective lens. The 2-time electrons 211 reflected from the sample are detected by the detector 209 and displayed as an image.

Examples of optical conditions for obtaining a high-resolution image include an acceleration voltage for accelerating electron beams from an electron source, an excitation current of a condenser lens or an objective lens, and a current and a voltage of a deflector. In addition, as an index of performance required for the inspection apparatus to obtain a high-resolution image, aberration and acceptance (acceptance) can be given. The aberration means that when the electron beam is converged by a lens or the like, the electron beam is not completely converged to one point, and when the value is large, distortion or blurring of the image is involved. The acceptability is a ratio of 2-time electrons collected in the detector, and when the value is small, a clear image cannot be obtained. Conventionally, after a wafer is obtained, an optimal optical condition is searched while looking at an image.

Fig. 3 shows a user use procedure and a process flow when the analysis result DB creating unit 109 creates the analysis result DB 104. Before obtaining a customer wafer, a user creates a model of the wafer based on the movement of the semiconductor device and customer information, analyzes the model, and creates an analysis result DB 104. First, an inspection apparatus model and a wafer model used for analysis are prepared and registered in the model DB101 (step 301). Next, IF102 sets a model, a focus point, and optical conditions in the analysis condition (step 302).

The user operates the analysis condition setting IF of fig. 4. Which is displayed on a display screen of the optimum optical condition extraction unit 110 or a display screen connected to the CD-SEM. The screen 401 for setting the analysis conditions IF is composed of a model setting unit 402, a focusing point selection unit 403, an optical condition input setting unit 404, and an analysis execution button 405. The model setting unit 402 inputs the marks of the inspection apparatus model and the wafer model registered in the model DB to select a model to be used for analysis. The focusing point selection unit 403 selects a focusing point previously associated with the wafer model such as the surface or the groove bottom. Alternatively, an arbitrary focus point may be input. The optical condition input setting unit 404 receives conditions such as an acceleration voltage, an electrode voltage, and an excitation current. It is also conceivable to analyze the combination conditions efficiently, input the maximum value, the minimum value, and the scale width of each condition, and automatically generate each combination condition. Then, by pressing an analysis execution button 405 provided on the screen, the analysis execution unit 103 executes analysis to obtain performance such as receptivity and aberration and an observation image. (step 303).

Next, an analysis result DB104 is created which is composed of analysis results in which models used for analysis, optical conditions as analysis conditions, and analysis results are associated with each other (step 304). Fig. 5 shows an example of the analysis result DB 104. In one line, 1 result is stored by integrating an inspection apparatus model and a wafer model used for analysis, a focusing point and an optical condition as input conditions of the analysis, output result performance of the analysis, and an image. When the model and the optical condition are different, a new line is added to store the result.

Fig. 6 shows the order of use by the user when the optimal optical condition is extracted by the optimal optical condition extraction unit 110 and the processing flow of the optimal optical condition extraction unit 110. In this process, the user operates the observation condition setting IF in fig. 7 as a use procedure, and the optimum optical condition extraction unit 110 performs a process corresponding to this operation. The screen 701 for setting IF using the observation conditions used in the procedure is composed of a wafer specification setting unit 702, a registered wafer model selecting unit 703, a focusing point selecting unit 704, an optimization index/priority setting unit 705, and an optimal condition extracting button 706. The observation condition setting IF is displayed on the display screen of the optimum optical condition extracting unit 110 or the display screen connected to the CD-SEM in the same manner as fig. 4. The focusing point selecting unit 704 has the same function as the focusing point selecting unit 403 shown in fig. 4.

The following describes the order of use and the order of processing. First, the user inputs specifications of the customer wafer, or selects a wafer close to the customer wafer from a wafer model registered in advance. The customer inputs the specification of the customer wafer by using the wafer specification setting unit 702. As for the specification of the customer wafer, the geometrical shapes such as the surface, cross-sectional image file, groove width, groove depth, and the like of the SEM image photographed at a low magnification are input as the specification of the wafer. These specifications are data transmitted from the client side via a network or a storage medium, and the optimum optical condition extraction unit 110 receives the data via an interface not shown. Further, a wafer provided as a standard wafer from a customer may be imaged in an inspection apparatus and used as a standard wafer.

The registered wafer model selection unit 703 is used, for example, to select a similar model according to the surface pattern shape of the wafer. The user selects a similar model by visually comparing the contents thereof, each model displayed in the registered wafer model selecting unit 703, and the image included in the specification.

Further, the user inputs a focusing point such as the surface or the bottom of the groove, that is, a focusing point indicating a position at which the inspection apparatus focuses on the sample, to the focusing point selecting unit 704.

That is, these inputs and selections are received by the observation condition setting IF105 of the optimum optical condition extraction unit 110 (step 601).

Next, the user inputs an index, a numerical value thereof, and a priority to be given priority when observing the wafer to the optimization index/priority setting unit 705. For example, the numerical values and priorities are targeted for the aberration and the receptivity. That is, the observation condition setting IF105 receives these inputs (step 602).

Then, when the user presses the optimal condition extraction button 706 provided on the screen, the wafer pattern search unit 106 of the optimal optical condition extraction unit 110 searches for a wafer pattern having a uniform focal point among wafers having similar patterns (step 603). The details of the search are as follows. First, an analysis result including a "wafer model" corresponding to the model selected by the registered wafer model selection unit 703 in step 601 is extracted once from the analysis result DB 104. Then, the image detected in the CD-SEM (observed image) is received, and an image similar thereto is extracted twice from the analysis result of the primary extraction, that is, a search is performed. That is, an image satisfying a predetermined criterion such as a similarity or the like having a predetermined degree or more is searched for by image search. This retrieval can be achieved by retrieving similar images using machine learning, for example.

In this embodiment, the extraction is performed once, but it may be omitted. In this case, the selection by the registered wafer model selector 703 and the input of the specification by the wafer specification setting unit 702 in step 601 may be omitted.

Next, the optimum optical conditions in the retrieved wafer are extracted (step 604). That is, the optical condition extraction unit 107 extracts the optical condition included in the analysis result of the image searched in step 603 as the optimal optical condition.

Here, a detailed process of step 604 of extracting optimal optical conditions in similar wafers is shown in fig. 8. This process is performed in the optical condition extraction unit 107. First, analysis results with the same focal point are extracted from the analysis result DB104 for similar wafers (step 801). That is, it is determined that the focus point including the data of the image retrieved in step 603 has the analysis result of the focus point matching the focus point input via the focus point selecting unit 704 in step 601. In addition, when the number of analysis results is a predetermined number (for example, 1), the following processing may be omitted, and the optical conditions included in the analysis results extracted in this step may be extracted to end the processing. In this case, the determined optical condition is output. The output destination may be a display screen of the optimal optical condition extraction unit 110, or may be a CD-SEM of fig. 2. In the case of the CD-SEM of FIG. 2, the optical conditions are accepted in the CD-SEM, which is followed to adjust the optical conditions.

Next, the performance index and priority set by IF105 for the observation condition are set to create the extraction condition of the analysis result (step 802). That is, in step 602, an extraction condition is created that satisfies the index input via the optimization index/priority setting unit 705 and the priority order of the numerical value.

Here, it is determined whether or not the priority is set in step 602 (step 803).

When the priority is set, a condition matching the priority is created and a result of satisfaction is extracted. For example, when the priority is 1 when the aberration is 1 μm or less and the priority is 2 when the receptivity is 0.8 or more, the conditions are made 2 for the aberration and the receptivity. That is, the optical condition that the aberration becomes 1 μm or less and the optical condition that the acceptance is 0.8 or more are extracted separately (step 804).

On the other hand, when the priority is not set, optical conditions satisfying the respective conditions are extracted using multi-objective optimization and machine learning (step 805). As for the multi-objective optimization, there is a method of sampling a representative analysis result, creating a response surface, and using a genetic algorithm, and as for the machine learning, there is a method of using a neural network, a decision tree, or the like.

The methods described so far are examples of extraction of optimal optical conditions, and the present invention is not limited to these methods.

Example 2

Next, fig. 9 shows an automatic inspection device adjustment system according to embodiment 2. In embodiment 2, the analysis result DB creation unit 109 connects various devices and databases via the network 901. Further, the specification can be received via the network 901, or the analysis result DB104 can be shared among a plurality of inspection apparatuses. That is, the analysis result DB creation unit 109 realizes clouding.

For example, consider the following case: even if the specifications of the wafers to be inspected are largely different after the delivery of the inspection apparatus, the optical conditions need to be set again, but the model DB does not include a model of the wafer. Therefore, the analysis result DB creation unit 109 of the inspection apparatus automatic adjustment system is built in a PC of an inspection apparatus manufacturer, and the optimum optical condition extraction unit 110 is provided in the inspection apparatus main body and connected to the inspection apparatus main body via a network 901 such as the internet. Thus, the analysis result DB104 can be accumulated in accordance with the latest state of the wafer and the inspection apparatus by updating the inspection apparatus model when the inspection apparatus is improved in addition to the wafer model of the model DB.

With the above configuration, it is possible to extract optimal optical conditions corresponding to modification of the inspection apparatus and updating of the wafer pattern after delivery.

Example 3

Fig. 10 shows another example of the extraction method of the optimal optical condition shown in fig. 8. This is an extraction procedure assuming optimal optical conditions for the case of measuring the width of the deep hole bottom. In order to measure the width with high accuracy, it is necessary to focus the wafer on the bottom, but the depth may vary in an actual wafer. The processing performed in the optical condition extraction section 107 for solving this problem is described below.

First, an estimated value of the variation in depth is set (step 1001). Which is set based on past performance values and information from the client. For example, a depth actual result value in the past examination input via the observation condition setting IF105 and a variation in depth included in information from the client are calculated and used as estimated values.

Next, optimum optical conditions obtained from analytical models (wafer models, inspection apparatus models) in which the depths are changed by several types are extracted in accordance with the estimated values (step 1002). This is achieved by extracting a plurality of analysis models from the analysis result DB104 within the range of the estimated values (deviations) to equalize the difference in depth.

The analysis result including the extracted analysis model is specified, and the optical conditions of the analysis result are set in the inspection apparatus to measure the pattern (step 1003).

Then, the optical condition with the best measurement result is selected (step 1004). This selection is achieved by selecting an optical condition that best satisfies a predetermined condition for the measurement result, or by the user selecting the measurement result from the screen displayed on the optimal optical condition extraction unit 110 or the like.

Finally, it is determined whether the measurement accuracy is sufficient (step 1005). If sufficient, the process is finished. If not, the depth of the analytical model of the best optical condition obtained in step 1004 is set to be the center, and the range of depth variation is narrowed (i.e., the process returns to step 1001). Then, the apparatus was set and measured. The procedure was repeated. The determination in step 1005 may be made based on whether or not a preset condition is satisfied, or the user may confirm the contents of the screen displayed on the optimum optical condition extraction unit 110 or the like.

By selecting the best optical condition in combination with the measurement result in the apparatus in this way, the optimal optical condition can be extracted even when the information on the specification of the customer wafer is small.

In the present embodiment, the description has been given by taking "the width of the bottom of the deep hole" as an example, but the present invention can be applied to the size of the inspection target including the sample of the wafer, and can also be applied to "the width of the bottom of the deep hole". The dimensions include the depth of the bottom of the deep hole. The present embodiment can be applied to any of the configurations shown in fig. 1 and 9.

Description of reference numerals

101.. model DB, 102.. analysis condition setting IF, 103.. analysis execution unit, 104.. analysis result DB, 105.. observation condition setting IF, 106.. wafer pattern retrieval unit, 107.. optical condition extraction unit, 108.. optical condition setting unit, 109.. analysis result DB preparation unit, 110.. optimum optical condition extraction unit, 201.. lens barrel, 202.. electron source, 203.. 1 converging lens, 204.. 2 converging lens, 205.. deflector, 206.. objective lens, 207.. sample, 208.. stage, 209.. detector, 210.. 1 electron, 211.. 2 electron, 401. analysis condition setting screen, 402.. model setting unit, 402.. 403.. focus point selection unit, 404.. input condition setting unit, 405.. analyze the execution button, 701.. observe the screen of the condition setting IF, 702.. wafer specification setting, 703.. register the wafer model selecting, 704.. focus point selecting, 705.. optimize the index/priority setting, 706 … optimum condition extraction button, 901.. network.