JP2001281178A - Defect detecting method, manufacturing method of semiconductor device, and defect detector - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a defect detecting method capable of accurately detecting a minute defect on a substrate by using a simple device. SOLUTION: Pattern data corresponding to a portion having a pattern formed thereon and background data corresponding to a background are detected from image data in an imaging area of an SEM(scanning electron microscope) 20. By performing calculation, based on the pattern data and the background data, an appropriate threshold can be set. The SEM image data is binarized by using the threshold to perform defect detection.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a defect detection method for detecting microscopic defects such as dents generated during the manufacture of a semiconductor device by using a scanning electron microscope, a semiconductor device manufacturing method, and a defect. It relates to a detection device.

[0002]

2. Description of the Related Art Conventionally, detection of minute defects generated during an LSI manufacturing process such as a film forming and etching process and measurement of the shape of the minute defects have been performed using an optical microscope.
Scanning electron microscopes (hereinafter referred to as SEMs) are being used with the miniaturization of semiconductor devices. As a method for detecting such minute defects, for example, there is a pattern matching method. In this method, a similarity is calculated for an image of a reference defect registered in advance with respect to an image to be measured while shifting the position over the entire screen, and an image whose value exceeds a set value is detected as a defect. The shape can be measured by counting the pixels and determining the area, length, and the like for each defect detected in this way. There is also a method of detecting defects by sequentially comparing images of the same design pattern. this is,
A method of comparing patterns having the same shape in different places and detecting a difference between the patterns as a defect, for example, extracting a defect such as a large foreign substance from a regular wiring pattern and measuring the shape. can do.

[0003]

The defects to be solved in the present invention are defects such as minute holes formed on the wafer, and have a small area in the image of the object to be measured. The shape is not constant. The method using pattern matching for such a minute defect is not suitable for extracting a defect having an irregular shape because the defect pattern needs to be registered in advance. Further, when a method of sequentially comparing images of the same design pattern to detect a defect is used, it is necessary that the defect occupies a certain size in the image to be measured in order to find a portion having a different regularity. Yes, it is difficult to extract minute defects. Further, both methods have a problem that when a general-purpose processor is used for processing, the processing time becomes longer. Further, a device using dedicated image processing hardware capable of high-speed processing has a problem that it is expensive. For these reasons, defects such as minute holes generated on the wafer have been visually checked to determine the presence or absence, and the shape has been evaluated, which is troublesome and cannot be quantitatively evaluated. There was a problem. SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and a micro-defect detection method, a micro-defect detection device, and a semiconductor device manufacturing method capable of detecting a micro-defect using a quick and simple device. The purpose is to provide.

[0004]

As a means for solving the above problems, according to the present invention, an electron beam is scanned on a sample surface on which a pattern is formed by using a scanning electron microscope. In a defect detection method for acquiring image data corresponding to the amount of electrons detected by scanning an electron beam and detecting a defect on the surface of the sample based on the image data, A first detecting step of detecting, as pattern data, the value of image data corresponding to a portion where the pattern is formed, and a portion of the image data other than the portion where the pattern is formed. Image data corresponding to the part
A second detection step of detecting the value of the data as background data;
A threshold value calculating step of calculating a threshold value based on the pattern data and the background data; and a step of selecting the image data based on the threshold value to detect a defect on the sample surface. And a defect detecting step. Even when an electron beam is scanned with respect to a minute defect having a concave shape on a sample surface, the amount of detected electrons is often small. Therefore, it is possible to detect minute defects by selecting image data obtained according to the amount of electrons at a predetermined threshold value. In addition, it is often difficult to appropriately set such a threshold value due to brightness or the like in the implementation conditions. The present invention
As a method for setting a threshold value, image data corresponding to a portion where a pattern is formed and image data corresponding to a background are detected, and a threshold value is determined based on the two image data. It is something to set. By adopting such a method, when setting the threshold value, it is possible to take in relative elements such as the difference between two image data, and it is possible to set an appropriate threshold value.

[0005] In addition, since the defect inspection usually inspects the pattern portion, the inspection using the image data corresponding to the portion where the pattern is formed increases the versatility of the defect detection method, There is also an effect that inspection can be performed easily and in a short time. Also,
Using a scanning electron microscope, an electron beam is scanned over the sample surface on which the pattern is formed, and image data corresponding to the amount of electrons detected by scanning the electron beam is acquired.
A defect detecting method for detecting a defect on the sample surface based on the image data; a step of detecting a maximum value of the image data; a step of detecting an average value of the image data; A threshold value calculating step of setting a threshold value based on a maximum value and the average value; and a defect detecting step of selecting the image data based on the threshold value and detecting a defect on the sample surface. This is a defect detection method characterized by comprising: Further, the defect detecting step includes a step of binarizing the image data by the threshold value to separate the image data into a defective area and a non-defective area, and measuring a shape of the defective area. 3. The defect detection method according to the above item 1 or 2, wherein

[0006] In the threshold value calculating step, the threshold value TLV is obtained by calculating the value of the pattern data by PLV.
When the value of the background data is BLV, TLV = BLV + (PLV-BL) using a predetermined coefficient RATE.
V) x RATE. Further, the defect detection method is characterized in that the pattern is a conductive substance. A step of forming a thin film on the semiconductor substrate; a step of forming a predetermined pattern on the thin film by lithography and etching; and an inspection for inspecting the pattern formed on the semiconductor substrate. A method of manufacturing a semiconductor device by forming a semiconductor element on the semiconductor substrate, wherein the patterning is performed by using a scanning electron microscope. Scanning the surface of the semiconductor substrate with an electron beam and acquiring image data corresponding to the amount of electrons detected by the scanning of the electron beam; A first detection step of detecting, as pattern data, the value of image data obtained by scanning a pattern, and scanning a portion of the image data other than the portion where the pattern is formed. Image obtained De - Background de values of data - second detected as data
Detecting the pattern data and the background data.
A threshold calculation step of setting a threshold based on data
A defect detection step of selecting the image data based on the threshold value and detecting a defect on the surface of the semiconductor substrate.

An electron beam is scanned over the surface of the sample on which the pattern is formed using a scanning electron microscope, and image data corresponding to the amount of electrons detected by the scanning of the electron beam. In a defect detecting apparatus for acquiring data and detecting a defect on the surface of the sample based on the image data, a value of the image data corresponding to a portion where the pattern is formed in the image data is obtained. First detecting means for detecting as pattern data, and detecting, as background data, values of image data corresponding to portions of the image data other than the portion where the pattern is formed. Second detecting means, threshold value calculating means for setting a threshold value based on the pattern data and the background data, and selecting the image data based on the threshold value. Defect detection means for detecting a defect on the surface of the sample, It is a defect detection apparatus according to claim. An electron beam scanning section for scanning an electron beam on the sample surface on which the pattern is formed; and an electron beam reflected by the sample surface or a second electron beam emitted from the sample surface by the scanning. An electron detector for detecting a next electron, image data obtaining means for obtaining image data based on the intensity of the electrons detected by the electron detector, and the pattern of the image data First detecting means for detecting the value of image data corresponding to the portion where the pattern is formed as pattern data, and corresponding to a portion of the image data other than the portion where the pattern is formed; Second detecting means for detecting the value of image data to be processed as background data, and threshold value calculating means for setting a threshold value based on the pattern data and the background data. , The image data is sorted by this threshold value and And defect detecting means for detecting a defect on a sample surface, a defect detection apparatus, characterized in that it comprises a.

[0008] Further, in the above defect detecting apparatus, the pattern is made of a conductive material.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of a defect detection method, a defect detection apparatus, and a semiconductor device manufacturing method according to the present invention will be described with reference to the drawings. (First Embodiment) FIG. 1 shows a configuration diagram of a defect detection apparatus 10 according to a first embodiment of the present invention. This defect detection apparatus 10 scans a sample W with an electron beam and detects the amount of electrons reflected and emitted from the surface of the sample W to obtain SEM 20 original image data. SEM 20), an image input device 11 for inputting SEM original image data, and an image input device 1
An image processing unit 12 for performing image processing on the SEM original image data output from 1 and a monitor 13 for displaying a processing result of the image processing unit 12. SEM2
Numeral 0 denotes an electron beam scanning unit 23 for scanning an electron beam on a sample W mounted on a stage 22 in a chamber 21, and for detecting reflected electrons and secondary electrons. An electronic detection unit (not shown) is provided. Data corresponding to the amount of detected electrons is stored as SEM original image data in, for example, an MO.

The SEM original image data input to the image input device 11 is output to the image processing unit 12. Image processing unit 1
Reference numeral 2 denotes a spatial filtering unit 15 for performing a filtering process on the SEM original image data as shown in FIG.
And pattern data based on the filtered data.
A pattern data / background data detector 16 for detecting data and background data; and a threshold for setting a threshold value based on the detected pattern data and background data. A value setting unit 17, a binarization processing unit 18 for binarizing data with a set threshold value, a shape measurement unit 19 for measuring a shape from the binarized data, Consists of The operation of defect detection in the above configuration will be described. FIG. 3 is a flowchart showing the operation of the defect detection. The description will be made in the following order. FIG. 4 shows that the electron beam is scanned on the surface of the sample W on which the aluminum wiring pattern 25 is formed using the electron beam scanning unit 23, and the resulting 8-bit gray scale SEM is obtained. The original image is shown (S1). The line width of the wiring is about 200 nm, and the magnification is about 80,000 times. FIG.
The black dots shown in FIG. 3 indicate the minute defects 2 formed on the surface of the sample W.
6. Since such SEM original image data often contains random noise, if necessary, a spatial filtering process is performed to smooth the image (S
2). The filtering matrix at this time includes, for example, a 3 × 3 filter coefficient (111 111 111),
(111 121 111) or the like is used. By performing this spatial filtering process, it is possible to suppress erroneous detection of defects and deterioration of shape measurement accuracy. FIG. 5 shows an SEM20 image after the filter processing.

Subsequently, pattern data and background data required for the threshold setting process are detected (S3). FIG. 6 shows image data of an image at each position on the straight line AA 'in FIG. In general, aluminum has a large electron reflectance and emission rate, so that the value of the image data corresponding to the wiring pattern 25 is large as is apparent from FIG. Further, since the minute defect 26 has a concave shape, the amount of detected electrons is small. Therefore, as shown in FIG. 6, the value of the image data corresponding to the minute defect 26 becomes small. Therefore, for example, if the value of the largest image data among all the pixels in the inspection area is detected, that image data is detected.
The value of the data can be the value of the pattern data. The value of the background data is the value of image data corresponding to a portion other than the portion where the pattern is formed. However, since the area of the pattern in the inspection area is not so large, for example, the value of The average value of the image data may be used as the value of the background data. The value PLV of the pattern data and the value BLV of the background data are shown in FIG. Subsequently, a threshold value setting process is performed based on the thus obtained pattern data PLV and background data BLV (S4). The contrast of the image is expressed by the difference between the pattern data PLV obtained above and the background data BLV so that a defect can be cut out even if the contrast and brightness of the SEM image change. BLV represents the brightness of an image. These values and a coefficient R for extracting a preset defect
Using ATE, the binarization threshold TLV is calculated as follows: TLV = B
LV + (PLV−BLV) × RATE (Equation 1) For example, when the TLV is calculated with RATE = −0.3, the threshold is set so as to cut out only the defective portion 4 as shown in FIG.

Then, binarization processing is performed using the threshold value obtained above (S5). That is, binarization processing is performed on the multi-valued and displayed image data depending on whether the value is larger or smaller than the threshold value TLV, and divided into a defective portion and a non-defective portion.
FIG. 7 shows the result of the binarization process. White portions are defective portions, and black portions are non-defect portions. If a binarized image including a defect portion and a background portion as shown in FIG. 7 is obtained, the shape of each defect can be measured (step S5). For this purpose, first, defective portions are selected one by one by a labeling process. For example, to find the area of a defect,
For each of the selected defective portions, the total number of pixels included in the defect is counted, and the sum thereof is multiplied by the area value per pixel. Further, in order to obtain the major axis and minor axis of the defect, for each of the selected patterns, the number of pixels of the pattern is counted in the x and y directions, and this is multiplied by the dimension value per pixel. I just need. In order to determine the perimeter, the number of pixels on the outermost periphery of each of the selected defective portions may be counted, and multiplied by the dimension value per pixel. The circularity can also be determined using the area and the perimeter determined as described above.

Through the above-described image processing steps, SE
It is possible to finally extract a desired defect from the M image and measure the shape of each defect. In particular, even when the defect is minute in the image of the measurement object, the minute defect is binarized by the SEM image using a threshold set based on the pattern data and the background data. Can be extracted without overlooking. For example, when the defect detection method of the present invention is described in C language and processed by a personal computer equipped with a general-purpose processor, the SEM image (620 × 400 pixels) of FIG. 4 is processed as shown in FIG. The area, minor axis, and major axis of each defect could be measured, and the defect could be quickly measured using a general-purpose processor. In particular, even in the case where the defect is minute in the image to be measured as shown in FIG. 4, the method of the present invention can extract the minute defect without overlooking it. In addition, the major and minor axes of each defect were measured several times for the same image while changing the conditions of focus and brightness.
A sufficient measurement reproducibility of 0.01 μm was obtained. As described in detail above, when the pattern is inspected using the SEM,
By detecting pattern data and background data from an EM image and setting a threshold value based on the data value, an appropriate threshold value for detecting a defect can be easily set. It becomes possible. Further, it is possible to easily detect the shape of the defective portion by performing a binarization process using the threshold value.

Although the defect detection apparatus according to the present embodiment includes the SEM, the defect detection apparatus may include an image input device and an image processing device without the SEM. Also,
The image input device may input an image using a hard disk, a floppy (registered trademark) disk, an image scanner, or the like. Alternatively, the SEM and the image input device may be connected using a cable to detect a defect in real time. In addition, it goes without saying that the present embodiment can be variously modified. (Second Embodiment) A second embodiment of the present invention will be described with reference to the drawings. FIG. 8 is an explanatory diagram schematically illustrating an example of a manufacturing process of a semiconductor device. For example, LSI
As shown in FIG. 8, a manufacturing process of forming a film on a semiconductor substrate 40, a process of applying a resist 42 on a surface on which the film is formed, and a process of irradiating ultraviolet rays to remove a pattern of a reticle M Exposure step of transferring to resist 42
Is developed using an organic solvent to form a resist pattern on the substrate, an etching step is performed to process a part of the formed film 41 by etching, and an ashing step is performed to remove the resist 42 after etching. And Then, a semiconductor element having a predetermined function is manufactured on the semiconductor substrate 40 by repeating the above steps.

After each of the above steps, an inspection step is often performed to evaluate the workmanship of the process. For example, the image in FIG. 4 is an electron micrograph at the time of inspection after the etching process. In the method of manufacturing a semiconductor device according to the present invention, inspection is performed in these inspection steps using the defect detection method according to the present invention. By performing the inspection in this manner, a minute defect can be detected quickly and accurately, so that it is possible to improve the throughput of manufacturing a semiconductor device. Note that the method of manufacturing a semiconductor device according to the present invention is also applicable to post-process inspections other than those described here. For example, the present invention can be applied to the inspection after the CMP process, and is not limited to the process described in this embodiment. Needless to say, the inspection method shown in the first embodiment is not used in all the inspection steps. In addition, the present embodiment can be variously modified.

[0015]

According to the present invention, pattern data corresponding to a portion where a pattern is formed and background data corresponding to a background are detected from the SEM image data, and the pattern data and the background are detected. A microdefect detection method, a microdefect detection device, and a semiconductor with improved throughput that can detect a microdefect using a quick and simple device by calculating a threshold value based on data A method for manufacturing the device can be provided.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a defect detection device according to a first embodiment.

FIG. 2 is a diagram showing an image processing unit of the defect detection device according to the first embodiment.

FIG. 3 is a flowchart of defect detection according to the first embodiment.

FIG. 4 is a view showing SEM original image data according to the first embodiment;

FIG. 5 is a diagram showing an image subjected to a filtering process in the first embodiment.

FIG. 6 is a diagram showing image data for each position in the first embodiment.

FIG. 7 is a diagram showing an image after a binarization process in the first embodiment.

FIG. 8 is a schematic view of a manufacturing process of the semiconductor device according to the second embodiment.

[Explanation of symbols]

11 image input device, 12 image processing unit, 23 electron beam scanning unit.

Claims (9)

[Claims] An electron beam is scanned over a sample surface on which a pattern is formed using a scanning electron microscope, and image data corresponding to the amount of electrons detected by scanning of the electron beam. In a defect detection method for acquiring data and detecting a defect on the sample surface based on the image data, the image data corresponding to a portion of the image data where the pattern is formed. A first detection step of detecting the value of the pattern data as pattern data; and a step of detecting a value of the image data corresponding to a portion of the image data other than the portion where the pattern is formed. A second detection step of detecting as data, a threshold value calculation step of calculating a threshold value based on the pattern data and the background data, A defect detection step of selecting the data and detecting defects on the sample surface; Defect detecting method, characterized in that it comprises. 2. A scanning electron microscope is used to scan an electron beam on the surface of a sample on which a pattern is formed, and image data corresponding to the amount of electrons detected by scanning the electron beam. A defect detection method for acquiring data and detecting a defect on the surface of the sample based on the image data; a step of detecting a maximum value of the image data; and a step of detecting an average value of the image data. A threshold value calculating step of setting a threshold value based on the maximum value and the average value; and a defect for selecting the image data based on the threshold value and detecting a defect on the sample surface. A defect detection method, comprising: 3. The defect detection step includes: a step of binarizing the image data according to the threshold value to separate the image data into a defect area and a non-defect area; and a step of measuring a shape of the defect area. The defect detection method according to claim 1, comprising: 4. In the threshold value calculating step, when the threshold value is represented by TLV, the value of the pattern data is represented by PLV, and the value of the background data is represented by BLV, 2. The defect detection method according to claim 1, wherein the threshold value TLV is calculated as TLV = BLV + (PLV-BLV) * RATE using a predetermined coefficient RATE. 5. The method according to claim 1, wherein the pattern is a conductive substance. 6. A step of forming a thin film on a semiconductor substrate, a step of forming a predetermined pattern on the thin film by lithography and etching, and an inspection of the pattern formed on the semiconductor substrate. A method of manufacturing a semiconductor device by manufacturing a semiconductor device by forming a semiconductor element on the semiconductor substrate, wherein the patterning is performed by using a scanning electron microscope. Scanning the surface of the formed semiconductor substrate with an electron beam, and acquiring image data corresponding to the amount of electrons detected by the scanning of the electron beam; A first detection step of detecting, as pattern data, the value of image data obtained by scanning a portion where the pattern is formed; and the pattern of the image data. Is formed A second detection step of detecting, as background data, the value of image data obtained by scanning a portion other than the portion, and a threshold based on the pattern data and the background data. A threshold value calculating step of setting a value; and a defect detecting step of selecting the image data based on the threshold value and detecting a defect on the surface of the semiconductor substrate. Method. 7. A scanning electron microscope is used to scan an electron beam on the surface of a sample on which a pattern is formed, and image data corresponding to the amount of electrons detected by scanning the electron beam. A defect detection device for acquiring data and detecting a defect on the surface of the sample based on the image data, wherein the image data corresponding to a portion of the image data where the pattern is formed; A first detecting means for detecting the value of the image data as pattern data; and a value of the image data corresponding to a portion other than the portion where the pattern is formed in the image data. Second detecting means for detecting as data, threshold calculating means for setting a threshold based on the pattern data and the background data, and image data based on the threshold. Defect detection means for selecting defects and detecting defects on the sample surface, Defect detection apparatus, characterized in that it comprises. 8. An electron beam scanning section for scanning an electron beam on a sample surface on which a pattern is formed, and reflected electrons reflected on the sample surface by the scanning or emitted from the sample surface. An electron detector for detecting the detected secondary electrons; image data obtaining means for obtaining image data based on the intensity of the electrons detected by the electron detector; First detecting means for detecting, as pattern data, the value of image data corresponding to a portion where the pattern is formed; and a portion of the image data other than the portion where the pattern is formed. Second detection means for detecting the value of image data corresponding to the portion as background data, and a threshold value for setting a threshold value based on the pattern data and the background data Calculating means, and the threshold value is used to calculate the image data. Defect detection means for selecting and detecting defects on the surface of the sample. 9. The defect detecting apparatus according to claim 7, wherein said pattern is made of a conductive material.