JP2000180377A - Apparatus for inspecting defect/foreign matter - Google Patents

Abstract

PROBLEM TO BE SOLVED: To inspect dust/foreign matters/shape defects distributed in three dimensions on a substrate over an entire height of a surface at high speed by arranging two or more optical systems of mutually different focal positions. SOLUTION: The apparatus for inspecting defects/foreign matters of patterns formed on a substrate such as a wafer 16 or the like has, for example, two series each consisting of an image input lens system 2 and a line sensor camera 1. A half mirror 18 and a total reflection mirror 19 are arranged on respective lens systems so that an equal point can be picked up. Images of the equal point which have different focal points are simultaneously input to the two line sensor cameras 1, A/D converted, and then input to pipe line process parts 7 and overlay process parts 8 of an equal number to the cameras. The image detected and processed by the pipe line process parts 7, a general-purpose processor 9 and an image-processing special processor 12 is displayed together with a compressed image as an output of the overlay process parts 8 on a monitor 15.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a dust / foreign matter inspection system for detecting and inspecting dust / foreign matter / shape defects of a pattern formed on a substrate at high speed.

[0002]

2. Description of the Related Art Conventionally, when performing a visual inspection of a patterned wafer, a target is regarded as a two-dimensional plane, a focus is fixed at a predetermined focal position at an arbitrary point on the wafer, and an image is input. However, dust, foreign matter, and shape defects are detected by cell comparison when the input image has a repetitive pattern, and by chip comparison when the input image is not a repetitive pattern.

However, the surface to be inspected of the semiconductor wafer has a three-dimensional shape microscopically, and the depth of focus in the conventional inspection method has a difference in height. -It was difficult to detect foreign matter and shape defects.

For this reason, Japanese Patent Application Laid-Open No. Hei 4-93752 discloses a technique in which image input is performed a plurality of times at different focal positions, and inspection is performed over the entire height of the wafer surface using a composite image obtained by integrating these images. Was.

[0005]

However, the above-mentioned prior art has the following problems. (A) Usually, in a wafer inspection apparatus, high resolution and high speed are required. Therefore, a line sensor is often used as an image input system. However, since scanning in the direction orthogonal to the line sensor is performed mechanically by a stage or the like, it takes a long time to input information on the inspection area as an image. In particular, when inputting images at a plurality of focal positions, mechanical scanning by a stage is performed at each focal position to input an image, so that the total inspection time is enormous.

(B) In the case of performing an inspection by combining images by integration, it is necessary to reproduce the mutual positional relationship at the time of inputting each image with high accuracy. However, it is not easy to reproduce the mechanical scanning start position with high accuracy and to keep the stage speed constant. Furthermore, if such conditions are not satisfied, a slight positional shift occurs, and the image becomes out of focus at the time of image synthesis, which greatly affects the inspection result.

[0007]

A defect according to claim 1 is provided.
A foreign substance inspection apparatus is an inspection apparatus that detects a shape defect, a scratch, or a foreign substance of a pattern formed on a substrate, and includes two or more optical systems having mutually different focal positions and an object captured by the optical systems. Imaging means for converting an image of an object into an electric signal;
Based on the image information obtained by the imaging means,
It is characterized by comprising an information processing section for detecting scratches / foreign matter.

According to a second aspect of the present invention, there is provided an inspection apparatus for inspecting a defect or a foreign matter on a repetitive pattern according to the first aspect of the present invention, wherein: means for correcting illumination unevenness with respect to an image input to a processing apparatus; Means for moving by a distance, spatial filter processing means for performing smoothing processing, maximum value filter processing means, minimum value filter processing means, image comparison means,
An image OR generating unit, an image gradation / gradation processing unit, an image binarization unit, an image contraction / expansion processing unit, an image compression, and a labeling unit, and a distance corresponding to a repetition pitch of the input image. , The result of performing the smoothing process is used as a reference image, and a difference between the reference image and the input image is detected.

According to a third aspect of the present invention, there is provided a defect / foreign matter inspection apparatus on a non-repeated pattern, wherein the non-repeated pattern inspection apparatus according to the first aspect is adapted to detect non-defective image data previously stored in a memory and an image inputted to a processing apparatus. Means for correcting illumination unevenness;
Spatial filter processing means for performing smoothing processing, maximum value filter processing means, minimum value filter processing means, image comparison means, image OR generation means, image density gradation processing means, image binarization means , Image compression / expansion processing means, image compression, and labeling means for detecting a difference between an image input from the image input unit and a non-defective image stored in a memory in advance.

[0010]

Embodiments of the present invention will be described below. FIG. 1 shows the configuration of this embodiment. In the figure, reference numeral 1 denotes a line sensor camera for inputting an image, and 2 denotes an image input lens system. In this embodiment, for simplicity, two lens systems are used, and an image input system having two types of focal points for the same field of view is provided. However, it is possible to arbitrarily add a lens system and a camera by adding the half mirror 18. The focus positions of the two lens systems of the present embodiment are shifted in advance, and the two lens systems are set so that the two optical systems can image different heights on the wafer surface.

A half mirror 18 and a total reflection mirror 19 are provided for each lens system so that two line sensor cameras can image the same place. The image input lens system, the half mirror 18 and the total reflection mirror 19 allow images having different focal positions at the same location to be formed.
Input to two line sensor cameras simultaneously.

The light beam emitted from the illumination light source 3 is bent downward by the half mirror 20, further bent by the total reflection mirror, and acts as the epi-illumination by the half mirror 18.

In the drawing, reference numeral 4 denotes a distance measuring sensor which measures the height of a reference position of the wafer 16 when the wafer 16 to be inspected is placed, and based on the measurement result, determines the initial height of the entire optical system. adjust.

By scanning the line sensor camera 1 and the XYθZ stage 5, the surface state of the wafer 16 to be inspected is
Input as an image that is dimensional information. The input image data is input to the line sensor camera interface 6 in parallel for each camera as time-series analog data,
After the analog-digital conversion, they are simultaneously input to the pipeline processing unit 7 and the overlay processing unit 8. The pipeline processing unit 7 and the overlay processing unit 8 are prepared by the same number as the number of cameras, and parallel processing for each camera is enabled.

The pipeline processing unit 7 performs image comparison by predetermined processing as described later, and transfers the image compressed to a predetermined size to the memory 10 of the general-purpose processor 9 via the input buffer 11. This process is performed by the pipeline processing unit 7 corresponding to each camera.

The memory 10 of the general-purpose processor 9 is connected to a processor 12 and a memory 13 dedicated to image processing and a local image bus 16 dedicated to image processing. The image compressed by the pipeline processing unit 7 is subjected to post-processing as needed by the general-purpose processor 9 and the image processing dedicated processor 12. In this post-processing, since the processing is performed on the compressed images, there is no need to perform parallel processing, and the respective images are sequentially processed.

The overlay processor 8 compresses the image. The compressed image is transferred to the memory 10 of the general-purpose processor 9 via the input buffer 11.

The pipeline processing unit 7 and the general-purpose processor 9
The image for which the detection processing has been completed by the image processing dedicated processor 12 is transferred to the control computer 14 together with the compressed image output from the overlay processing unit 8 and displayed on the monitor 15. Thus, the detected image is superimposed on the overlay image obtained by compressing the original image and displayed on the monitor, so that it is easy to visually confirm the defect position.

The detected defect position, size, etc. are transferred to the control computer as numerical data and stored.

Next, the processing flow of the pipeline processing unit 7, the general-purpose processor 9, and the image processing dedicated processor 12 will be described. These cooperate to detect dust, foreign matter, and shape defects. FIG. 2 shows a flowchart of the process. The flowchart of FIG. 2 is applied to the inspection target of the repetitive pattern.

In process 100, line sensor camera 1
And an XYθZ stage 5 to input an inspection image. The processing 101 corrects illumination unevenness. Image shift processing 1 by repetition pattern pitch for cell comparison
02, and a reference image generation process 103 for comparison is performed.

The smoothing process 104, 105 is performed for each of the original image and the reference image. Reference image generation processing 10
For No. 3, maximum value filter processing 106 and minimum value filter processing 107 are performed to prevent a pseudo defect at the time of image comparison. By comparing with the original image, an extraction process 108 for a portion whose density is higher than the maximum value filter image and an extraction process 109 for a portion whose density is lower than the minimum value filter image are performed. Through the above processing, a portion that is a candidate for dust, foreign matter, and shape defect is extracted.

Next, a sum image is formed by the inter-image operation processing 110 using the images of the respective processing results. The obtained sum image is subjected to a density conversion process 111 for cutting out a portion below a preset shading threshold, and then to a binarization process 112.

The binary image is subjected to a contraction / expansion process 113 for eliminating pseudo defects and a compression process 114 for image data.

The above processing is performed by the pipeline processing unit 7 in synchronization with image input, and is performed in parallel by each pipeline processing unit corresponding to each line sensor camera.

Since the data amount is greatly reduced by the image compression processing 114 (for example, when the data is compressed to 1/8 in the vertical direction and 1/8 in the horizontal direction, the data amount is reduced to 1/64). Post-processing can be easily performed in a single hour. Accordingly, the following processing is performed by the general-purpose processor 9 and the image processing dedicated processor 12.

After the number and the coordinate position of the compressed image are measured by a labeling process 115, an area threshold process is performed to leave a graphic having a size larger than a predetermined area.
Perform Step 6. By this processing, dust, foreign matter, and shape defects of a certain size or more are extracted. The above processing is sequentially performed for each compressed image transferred from the pipeline processing unit 7.

Next, a procedure for inspecting a non-repeated pattern will be described with reference to FIG. In this case, the non-defective image is stored in advance in the memory of the pipeline processing unit 7 as a reference image, and the same processing as that of the repetitive pattern is performed on this image and the input image.

Finally, processes 100 to 114 and process 20
FIG. 4 shows an embodiment of the configuration of the pipeline processing unit 7 for performing 0 to 214.

Since each processor uses a data flow type, it does not require access to a memory and constitutes a processing system only by the coupling shown in the figure.

Processing 100 to 114 and 200 to 21
Numeral 4 is assigned to nine processors 21 to 29, and is uploaded to each processor immediately before the inspection is performed. That is, in the case of the repetitive pattern, the processes 100 to 114 are uploaded, and in the case of not the repetitive pattern, the processes 200 to 214 are uploaded.

The processing items corresponding to each processor are as follows. Reference numeral 20 denotes a data input unit functioning as a router. The processors 1 of 21 are assigned processes 101 to 103 in the case of a repetitive pattern, and the processes 201 to 203 in the case of a non-repeated pattern.

Reference numeral 30 denotes an input buffer 11 of the general-purpose processor 9 for converting the data compressed by the processes 114 and 214 into data.
A shared memory 31 for transferring data to the transfer interface 31 is a transfer interface.

The labeling process was performed by the image processing dedicated processor 12 and the general-purpose processor 9.

Although the illumination light source 3 is used as a single light source,
It is also possible to use individual lighting for each camera. Further, it is also possible to omit table scanning by using an area sensor instead of a line sensor as an image input device. Although the present embodiment is directed to wafer inspection, the present invention can be applied to liquid crystal panel inspection.

[0036]

(1) According to the defect / foreign matter inspection method according to the first aspect, dust, foreign matter, and shape defects distributed three-dimensionally on a substrate having a repetitive pattern can be processed at high speed over the entire surface height. Inspection is possible.

(2) According to the defect / foreign matter inspection system according to the second aspect, dust, foreign matter, and shape defects of a pattern on a substrate having a non-repeated pattern and having a three-dimensional shape are distributed over the entire surface height. Inspection at high speed.

(3) According to a pipeline configuration using the data flow type processor according to claim 3, claim 1 is provided.
The data processing unit in the high-speed inspection according to the second aspect can be realized with a simple configuration.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram.

FIG. 2 is a flowchart of a repeated pattern inspection process.

FIG. 3 is a flowchart of an inspection process for a pattern that is not repeated;

FIG. 4 is a configuration diagram of a pipeline processing unit by a data flow type processor.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Line sensor camera 2 Image input lens system 3 Illumination light source 4 Distance measuring sensor 5 XYZθ stage 6 Line sensor camera interface unit 7 Pipeline processing unit 8 Overlay processing unit 9 General-purpose processor 10 Memory 11 Input buffer 12 Processor dedicated to image processing 13 Memory 14 control computer 15 monitor 16 wafer to be inspected 17 local image bus 18 half mirror 19 total reflection mirror 20 half mirror

Claims (3)

[Claims] An inspection apparatus for detecting a shape defect, a scratch, or a foreign matter of a pattern formed on a substrate, comprising: two or more optical systems having mutually different focal positions; and an object captured by the optical systems. A defect / contamination inspection apparatus comprising: an image pickup unit for converting an image of an object into an electric signal; and an information processing unit for detecting a defect, a scratch, or a foreign object based on image information obtained by the image pickup unit. 2. The inspection apparatus according to claim 1, wherein: means for correcting illumination unevenness with respect to the image input to the processing apparatus; means for moving the image by a certain distance; and spatial filter processing means for performing smoothing processing. Image processing means, maximum value filter processing means, minimum value filter processing means, image comparison means, image logical sum generation means, image density gradation processing means, image 2
The image processing apparatus further includes a value conversion unit, an image contraction / expansion processing unit, an image compression unit, and a labeling unit.The input image is repeatedly moved by a distance corresponding to the pitch, and a result of performing the smoothing process is set as a reference image. A defect / foreign matter inspection apparatus on a repetitive pattern, which detects a difference between an image and an input image. 3. The inspection apparatus according to claim 1, wherein the non-defective image data stored in a memory in advance, a means for correcting illumination unevenness with respect to an image input to the processing device, and a spatial filter processing means for performing a smoothing process. A maximum value filter processing unit, a minimum value filter processing unit, an image comparison unit, an image OR generation unit, an image density gradation processing unit, an image binarization unit, an image contraction / expansion processing unit, , Image compression,
A defect / foreign matter inspection apparatus on a non-repeated pattern, comprising: a labeling unit, for detecting a difference between an image input from an image input unit and a non-defective image stored in a memory in advance.