JP4857155B2 - Data processing apparatus, inspection system, and data processing method - Google Patents

Description

  The present invention relates to a data processing apparatus, an inspection system, and a data processing method for processing data observed and observed on a surface such as a semiconductor wafer, a photomask, a magnetic disk, and a liquid crystal substrate.

  In a semiconductor device manufacturing process, foreign matters and pattern defects (hereinafter referred to as defects) on the surface of a semiconductor wafer cause product defects. Therefore, after detecting defects with an inspection device, observe with a review device, digitize the size and type, store them in a database, analyze each batch of wafers, production lots, etc. Is constantly monitoring for problems. In addition, based on the observation result of the review device, the operator selects conditions under which similar defects can be easily extracted, and the conditions are set in the review device to improve the efficiency of defect observation.

  However, due to the increase in the number of detected defects due to the finer circuit patterns of semiconductor wafers, the multilayer structure, and the increased detection sensitivity of the inspection equipment, the number and types of analysis objects increase rapidly, It takes time to formulate countermeasures, which has become an obstacle to yield improvement. In addition, even if there is a defect, there are those that affect the performance of the final semiconductor device, and those that do not affect it. The former is called a fatal defect. Whether or not it can be judged is becoming important.

The above-described defect analysis work, that is, observation and classification, was performed by human eyes using images of an optical microscope and an electron microscope, so there is a difference in the technical level and experience of the operator to observe, There is a possibility that the classification of the type of defect to be observed is biased. In recent years, in order to solve this problem, automatic review (ADR: Automatic Defect Review) or automatic defect classification in which the apparatus automatically determines the size, shape, type, etc. of defects using image processing technology.
(ADC: Automatic Defect Classification) technology has begun to be introduced. For example, when observing a defective portion of a part from which a defect has been extracted by inspection (for example, a defective formation of a pattern formed on a wafer) using a SEM (Scanning Electron Microscopy) review device, the load on the operator is reduced. A system for efficiently performing work while reducing the amount has been proposed (for example, see Patent Document 1).

  In addition, with the miniaturization of semiconductor device processing dimensions, the number of fine defects has increased, the inspection conditions of inspection equipment that extracts defects are changed, and multiple defects extracted under each condition can be Since the number of output defects is increased by outputting, the amount of data handled by the inspection apparatus and the review apparatus is increasing. In addition, with the increase in sensitivity of the inspection apparatus, non-true defects may be extracted, and a lot of time is spent selecting true defects from a huge number of defect data.

As a countermeasure, RDC (Real-Time Defect) that classifies defects in addition to the inspection function.
An inspection device equipped with a (Classification) function has appeared, and automatic sorting of non-true defects has been attempted. The teaching data is what is clearly not a true defect from many defect images, and if there are similar images among other defect images, they are automatically classified into this category. is there. However, in order to improve the accuracy of judgment when classifying as a non-true defect, it is necessary to optimize the defect detection conditions of the inspection device and the conditions when acquiring images with the review device. An apparatus that handles a lot of information from an apparatus or a review apparatus and assists defect classification has been proposed (for example, see Patent Document 2).

  Also, in the semiconductor manufacturing process, multiple high-sensitivity inspection devices are owned and managed, but since the same process is inspected by multiple inspection devices, the same type of device can detect defects. There are devices with different sensitivities, and when inspections are performed, there are differences between the devices in the coordinates, number, defect size, etc. of the extracted defects, and a huge amount of time is spent reviewing and organizing the defect data considering these differences. Has been. Therefore, there is a system that supports the worker by connecting the data processing device that manages the information on the appearance inspection device and the review device through a communication line, and displaying the data transmitted from each device on the display of the data processing device. It has been proposed (see, for example, Patent Document 3).

JP-A-10-135288 JP 2001-156141 A (FIG. 2) JP 2006-173589 A

  As described above, in recent years, the number of defects detected by the inspection apparatus has increased, and the defect data includes those that are not true defects, so that important defects that require investigation and countermeasures for their occurrence are buried. There is a problem.

  In addition, even if there is a need to detect defects under the same inspection conditions in a plurality of inspection apparatuses, the detection conditions for the results obtained are different because the capabilities and sensitivities differ between inspection apparatuses, and similar defects are different. There is a possibility of being classified into categories.

  The present invention facilitates the processing of data obtained by automatically matching the defect information obtained by the inspection apparatus and the defect information obtained by the review apparatus, and can be easily classified for each type of defect. An object is to provide a data processing device, an inspection system, and a data processing method.

  In order to solve the above-mentioned object, an embodiment of the present invention provides a display for displaying inspection information relating to a defect of a sample transmitted from an inspection apparatus and review information obtained by observing the defect transmitted from a review apparatus. And a microprocessor that matches the inspection information and the review information to display the inspection information and the review information on the same defect side by side on the display, the microprocessor being selected from the defects Using multiple defect IDs as one axis, and displaying the feature quantities as a table on the display, with multiple conditions as one axis when the feature quantities included in the inspection information and review information displayed on the display are obtained In addition, the selected feature value data can be pasted to another software.

  According to the present invention, the defect information obtained by the inspection device and the defect information obtained by the review device are automatically collated with each other to facilitate the processing of the data, and the classification for each type of defect is possible. There is an effect that an easy data processing device, inspection system, and data processing method can be obtained.

  The overall configuration of the present invention will be described with reference to FIGS. Here, an example in which the present invention is applied to a semiconductor production line is shown. 1 and 2 are configuration diagrams showing the system configuration of the inspection work support system, and FIG. 3 is a screen diagram showing an example of an image displayed on the screen of the data processing apparatus.

  In FIG. 1, a semiconductor wafer manufacturing process 11 is usually in a clean room 10 in which a clean environment is maintained. In the clean room 10, an appearance inspection apparatus 1 that detects the appearance defect of a semiconductor wafer as a product 1, a review apparatus 2 that performs observation of an appearance defect based on data from the appearance inspection apparatus 1, that is, a review, at the end of the manufacturing process, A probe inspection device 12 for confirming electrical characteristics as a semiconductor is installed. The appearance inspection device 1 and the review device 2 are connected via a communication line 4 to a data processing device 3 for transferring image data from the inspection device. Since the data processing device 3 does not necessarily have to be in the clean room, it is usually installed outside the clean room 10, and a person other than the worker of the appearance inspection device 1 or the review device 2 can analyze data or the like. it can. Semiconductor wafers flow through a manufacturing process 11 in units of lots, and an appearance inspection process is incorporated in the middle of the manufacturing process 11. When the processing in the manufacturing process in which the appearance inspection is determined in advance is completed, the semiconductor wafer is carried to the appearance inspection apparatus 1 by an operator or a transfer machine (not shown), and the inspection is performed.

  In the inspection of the semiconductor wafer by the appearance inspection apparatus 1, defects are extracted by image comparison. One method is to store in advance an image of a normal circuit pattern that has the same shape as the circuit pattern to be inspected and has no defects, aligns it with the inspection image acquired for inspection, and then compares the difference by comparison. In this case, portions having different brightness can be extracted, and this is determined as a defect, and an inspection image is stored. The other method is a method in which an inspection image is used as a reference image. Two inspection images are compared, and if there is a difference, either inspection image has a defect. By repeating the process of comparing the second image and the third image, the process of comparing the third image and the fourth image, it is possible to determine which inspection image has a defect. The size of the range acquired as the inspection image is determined in consideration of the performance of the appearance inspection apparatus 1 such as a die unit which is the size of the semiconductor chip and a cell unit which is a repeated pattern provided in the die.

  In the appearance inspection apparatus 1, the extracted defect image and the coordinates of the defect in the appearance inspection apparatus 1 are sent as inspection data to the data processing apparatus 3 via the communication line 4 and stored in the data processing apparatus 3. As described above, when the RDC function is used in order to reduce the noise of defects due to the miniaturization of semiconductor devices and the performance improvement of the appearance inspection apparatus 1, the extracted defect classification results are also sent as inspection data.

  The wafer taken out from the appearance inspection apparatus 1 is sent to the review apparatus 2, and observation and analysis are performed on defects determined to require investigation of the cause of occurrence in the inspection data. As for the position of the defect, since the coordinate data of the defect is sent to the review device 2 among the data sent from the appearance inspection device 1 to the data processing device 3, the defect can be found based on the coordinate data. it can. The result of observing and analyzing defects in the review device 2 is sent to the data processing device 3 via the communication line 4 and stored.

In FIG. 2, the appearance inspection apparatus 1 detects minute defects on the wafer and sends inspection data 21 to the data processing apparatus 3 via the communication line 4. In addition to the defect image, defect ID, and coordinates in the appearance inspection apparatus 1, the inspection data 21 includes a product ID, a lot ID, a wafer ID, a die layout name, a manufacturing process, and an inspection, which are the types of wafers to be inspected. The date and time, the ID of the appearance inspection apparatus 1 and the like are included and stored in the data processing apparatus 3. The review device 2 shown in FIG. 1 includes types such as an optical review device 24 and an SEM review device 25, which are properly used depending on the contents of the inspection data 21. Here, SEM means Scanning Electron
Abbreviation for Microscope. Both the optical review device 24 and the SEM review device 25 are connected to the communication line 4 to transmit / receive data 22a, 22b, 23a, 23b to / from the data processing device 3.

  FIG. 3 is a screen diagram illustrating an example of an image displayed on the screen of the data processing device 3, and an example of the inspection data 21 is displayed. The example of the inspection data 21 includes a product ID, a lot ID, a wafer ID, a die layout name, a defect ID of a detected defect, coordinate data, a maximum gray level difference of the defect, a defect size, and the like. The definition of the maximum gray level difference will be described later. Although not shown in FIG. 3, the inspection data 21 includes, in addition to the above, an image of a defect detected by the appearance inspection apparatus 1, defect feature amount data as a result of RDC by the appearance inspection apparatus 1, and the like. . The format shown in FIG. 3 is not limited to this, and the format can be defined by selecting the data to be displayed as necessary.

  FIG. 4 is a screen diagram showing an example of an image displayed on the screen of the data processing device 3, and an example of items of defect feature quantity data is displayed with the name of an RDC parameter table. The defect feature data is transmitted from the appearance inspection apparatus 1 shown in FIG. 1 to the data processing apparatus 3 together with other defect data. If the defect data other than the defect image is text data in a predetermined format, the data capacity can be reduced.

  The defect feature amount shown in FIG. 4 will be described below. The maximum gray level difference is the absolute value of the brightness of the defect area of the difference image obtained from the image determined as a defect and its reference image. The reference image average gray level is an average value of brightness of an area corresponding to a defect area of the reference image. The defect image average gray level is the average value of the brightness of the defect area of the defect image. Polarity indicates whether the defect image average gray level is brighter or darker than the corresponding reference image average gray level, and if the defect image average gray level is lighter than the reference image average gray level, that is, “+”, it should be smaller. “−”. The inspection mode indicates an image comparison method used when a defect is detected, and includes die comparison, cell comparison, and mixed comparison thereof. The defect size and the number of defective pixels are values indicating the size of the detected defect. When the defect is elongated, the defect size width, the defect size height, and the defect size ratio are also used. The unit of the defect size is a unit of length such as micrometer, and the unit of the number of defective pixels is a pixel. The defect size ratio is a parameter representing the ratio between the width and height of the defect size, and is expressed as 1 if the width and height are the same, and 2 if the width is twice the height. The defective pixel differential value in the defective image is a differential value of the brightness of the pixel of the defective portion of the defective inspection image, and indicates the degree of change in gray scale. The defective pixel differential value in the reference image is a differential value of the brightness of the pixel in the reference image region corresponding to the defective region in the inspection image having a defect, and indicates the degree of change in gray scale. .

When observing a defect on a semiconductor wafer that has been inspected, the defect is carried to the review apparatus 2 shown in FIG. 1 for observation. The review device 2 acquires the inspection data 21 from the data processing device 3 using the product ID, lot ID, and wafer ID of the semiconductor wafer shown in FIG. 3 as key information.

  At this time, since the amount and contents of the inspection data 21 sent from the appearance inspection apparatus 1 to the data processing apparatus 3 are enormous, the data selected by the plurality of filter functions of the data processing apparatus 3 and selected from the inspection data 21 22 b is sent to the optical review device 24 and data 23 b is sent to the SEM review device 25 through the communication line 4.

  Based on the transmitted data 22b and 23b, a defect image is acquired by the optical review device 24 or the SEM review device 25, and the defect classification is performed using the ADC function installed in each review device. Is called. The result of defect classification is sent to the data processing device 3 through the communication line 4 as data 22a and 23a.

  FIG. 5 is a screen diagram illustrating an example of data screen display. In the data processing device 3, a computer (not shown) receives the inspection data 21 transmitted from the appearance inspection device 1 and the data 22 a transmitted from the optical review device 24 or the data 23 a transmitted from the SEM review device 25. The data is displayed on a display of the data processing device 3 (not shown) so as to be summarized in one table.

  The screen 30 displays a list and command buttons. The list includes a defect coordinate display field 34 for displaying the defect coordinates X and coordinates Y transmitted from the appearance inspection apparatus 1 shown in FIG. A defect image display field 35 for displaying an image and a defect feature value display field 38 for displaying a feature value of the defect are included. The defect feature amount display column 38 includes a condition display column 33 that indicates a condition when the feature amount of the defect is acquired, and a feature amount display column 31 that indicates the feature amount acquired based on the condition. When the amount of data is large, an area not displayed on the screen can be displayed with the scroll bar 47.

  Even if the condition for extracting the feature quantity of the defect is changed, if the coordinates of the defect are the same, it is regarded as the same defect, and a plurality of feature quantities are displayed side by side in the same row. For example, in FIG. 5, feature quantities A11, B11, and C11 are displayed side by side for a plurality of defects at coordinates X1 and Y1. In the matching result display field 32, the condition 1 for acquiring the feature amount is displayed as, for example, a circle mark, and the conditions 2 and 3 that are not acquired are displayed as, for example, a cross mark. The defect whose feature value has been acquired in this way is given the number shown in the matching ID display field 39. The order in which numbers are assigned is arbitrarily determined, but in this embodiment, the numbers are assigned in the order in which defect images are acquired.

  By attaching a check mark to the display target column 46 and instructing the sampling button 48, it is possible to select a defect for each matching ID and display only the selected defect. In the case of the filtering button 49, a defect with a check mark can be excluded from the display target.

  FIG. 6 is a screen diagram illustrating an example of data screen display. An example in which the condition is from condition 1 to condition 7 for one feature amount D is shown. Data 50 of the feature quantity D displayed on the screen can be selected by dragging with a pointer 51 such as a mouse and copied.

FIG. 7 is a screen view showing a list by a general spreadsheet software. For simplification of description, an example in which Condition 1 and Condition 2 are pasted from the data of the feature amount D in FIG. 6 is shown. By performing data processing such as graph display on the data in the region 71, the feature amount D can be displayed in a visual form instead of a number, so that efficient analysis can be performed.

  FIG. 8 is a screen diagram showing an example of the data screen display, and shows an example in which a scatter diagram is created from the table shown in FIG. FIG. 8 can be created by selecting the area 71 in the two vertical columns shown in FIG. 7 and instructing the spreadsheet software to create a graph with conditions 1 and 2 as axes. For example, assuming that the feature amount D is the size of a defect, and the condition 1 is the appearance inspection apparatus No. 1 and the condition 2 is the appearance inspection apparatus No. 2 and the data distribution is viewed, the defect detection performance of the apparatus If the effect of the difference is small, the data should be distributed side by side at an angle of 45 degrees for each axis. In the example shown in FIG. 8, since the data distribution is inclined toward the condition 1, the appearance inspection apparatus No. 1 tends to be detected as data having a larger defect size than the No. 2 for the same defect. I understand that. In addition, the degree of data variation can be easily grasped visually. In this way, it is possible to confirm the performance of the appearance inspection device and the compatibility between a plurality of devices.

  As described above, according to the present embodiment, the display for displaying the inspection information related to the defect of the sample transmitted from the inspection apparatus and the review information obtained by observing the defect transmitted from the review apparatus, and the inspection A data processing apparatus comprising a microprocessor that collates information and review information and displays inspection information and review information on the same defect side by side on a display, wherein the microprocessor has a plurality of defects selected from the defects With the ID of the one axis as one axis, and a plurality of conditions when the feature quantity included in the inspection information and review information displayed on the display is obtained as one axis, the feature quantity is displayed as a table on the display, and Since the feature value data selected from the feature values can be pasted to another software, the inspection device and the review device can From the result of automatically associating data on defect characteristics from each, the data characteristics can be visually confirmed by pasting the data into general spreadsheet software and creating a graph. Therefore, it is possible to easily find a clue for investigating the cause of the occurrence of a defect.

  In addition, the present embodiment provides an inspection system in which the data processing device, the inspection device, and the review device are connected by a communication line.

  Further, the present embodiment provides a data processing method executed by the data processing apparatus or inspection system.

The lineblock diagram showing the system configuration of an inspection work support system. The lineblock diagram showing the system configuration of an inspection work support system. The screen figure which shows an example of the image displayed on the screen of a data processor. The screen figure which shows an example of the image displayed on the screen of a data processor. The screen figure showing an example of the screen display of data. The screen figure showing an example of the screen display of data. The screen figure which displayed the list by general spreadsheet software. The screen figure showing an example of the screen display of data.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Appearance inspection apparatus 2 Review apparatus 3 Data processing apparatus 4 Communication line 10 Clean room 11 Manufacturing process 21 Inspection data 22a, 22b, 23a, 23b Data 24 Optical review apparatus 25 SEM review apparatus

Claims (6)

A display for displaying inspection information on the defect of the sample transmitted from the inspection apparatus, and review information obtained by observing the defect transmitted from the review apparatus;
A data processing apparatus comprising a microprocessor that matches the inspection information with the review information and displays the inspection information and the review information on the same defect side by side on the display,
The microprocessor uses, as one axis, IDs of a plurality of defects selected from the defects, and sets a plurality of conditions when the feature amounts included in the inspection information and the review information displayed on the display are obtained. Data processing characterized in that the feature quantity is displayed as a table on the display as one axis, and the feature quantity data selected from the plurality of feature quantities displayed can be pasted to another software apparatus.   The data processing apparatus according to claim 1, wherein the display displays a graph from the pasted feature amount data. An inspection apparatus that extracts a defect of a sample, assigns a defect ID to each defect, and acquires inspection information for each defect ID;
A data processing device that receives inspection information for each defect ID transmitted from the inspection device via a communication line, and generates at least the position information of the defect for each defect ID from the inspection information;
Based on the position information of the defect transmitted from the data processing device, the defect extracted by the inspection device is found and the defect is imaged, and the feature amount of the defect is determined from the imaged defect image of the defect. An inspection system including a review device that transmits the defect image and the feature amount to the data processing device for each defect ID.
The data processing apparatus causes the display to display the inspection information, the defect image, and the feature amount side by side with respect to the defect having the same defect ID, and the feature amount of the defect selected from the defect ID. An inspection system which can be copied to a spreadsheet software displayed on the display.   4. The data processing apparatus according to claim 3, wherein the display displays a graph from the copied feature quantity. The inspection information related to the defect of the sample transmitted from the inspection apparatus and the review information obtained by observing the defect transmitted from the review apparatus are matched, and the inspection information and the review information regarding the same defect are arranged side by side to the display. Display
The IDs of a plurality of defects selected from the defects are set as one axis, and the feature values when the feature values included in the review information are obtained are set as one axis, and the feature values are displayed as a table on the display. ,
Further, a data processing method characterized in that data of a feature quantity selected from a plurality of displayed feature quantities can be pasted to another software.   6. The data processing method according to claim 5, wherein a graph is displayed from the pasted feature amount data.

Images