JP3752338B2 - Macro inspection apparatus and process monitoring method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a macro inspection apparatus and a process monitoring method for inspecting a surface defect of a subject in a production line or inspection line of a wafer or a liquid crystal glass substrate.
[0002]
[Prior art]
Non-uniform film thickness of resist applied to the substrate surface or adhesion of dust in flat panel such as liquid crystal panel, PDP (plasma display panel), FED (field emission display), etc. These defects appear as pattern line width defects after etching and pinholes in the pattern.
[0003]
In order to prevent such defects, a macro inspection and a micro inspection for inspecting a substrate before etching for defects are conventionally performed. Micro inspection is micro defect inspection and pattern inspection using a device with a resolution of about 5 μm. On the other hand, macro inspection is to detect defects of relatively large area such as film thickness unevenness, exposure failure, and resist coating failure. This is a targeted examination. In the macro inspection, thin unevenness that cannot be found in the micro inspection and defects due to continuous pattern deviation are detected. Moreover, since a wide visual field can be inspected collectively, there is an advantage that inspection can be performed in a short time.
[0004]
Currently, visual inspection is still the mainstream for macro inspection, but automatic inspection has also been performed. In this automatic inspection method, there is one in which image processing is performed for each substrate to be inspected to make a pass / fail determination, and Japanese Patent Application Laid- Open No. 07-027709 discloses such a technique.
[0005]
Further, regarding the defect contents detected by the inspection apparatus, Japanese Patent Application No. 06-157424 and Japanese Patent Application No. 06-234818 disclose a technique capable of discriminating defect types to some extent.
[0006]
[Problems to be solved by the invention]
However, the above macro inspection apparatus only makes a pass / fail determination for each sheet.
In this way, not only the pass / fail judgment of each sheet, but also the defect occurrence status of the substrate that flows continuously and the change in the macro image of the non-defective substrate are confirmed, so that abnormalities and modulation of the manufacturing process can be performed before the defect occurs. There is a request to detect.
[0007]
Moreover, in the conventional macro inspection apparatus, the defect determination is performed for each substrate to be inspected, but the defect analysis is not performed in units of lots (about 20 to 30) or more. Therefore, it is difficult to detect common defects between lots or defects that are thin and cannot be detected by a single substrate, and as a result of overlooking such defects, noticeable defects eventually occur and cause a decrease in yield. It was.
[0008]
That is, in the above case, since the inspection image of the non-defective substrate does not remain in the inspection apparatus, it is impossible to observe how the inspection image changes. Therefore, until the NG substrate flows, it is impossible to predict an abnormality in the manufacturing process, and many abnormal substrates flow during that time.
[0009]
Under such circumstances, it is desired to realize an (inline) process monitor for predicting the occurrence of defects / abnormalities in the production line based on the information of the macro inspection. Here, the process monitor refers to a manufacturing management system that monitors a manufacturing process using defect information obtained from a defect inspection apparatus.
[0010]
The present invention has been made in consideration of such circumstances, and detects defects that cannot be detected on a single sheet basis, or statistically captures subtle modulation of the process even when only non-defective substrates are flowing. An object of the present invention is to provide a macro inspection apparatus and a process monitoring method capable of predicting the occurrence of a defect in advance and avoiding the occurrence of the defect.
[0011]
[Means for Solving the Problems]
The essence of the present invention is to add and accumulate macro images to be inspected flowing from the manufacturing process to create an added accumulated image, and analyze the added accumulated image to statistically capture subtle modulation of the process, and to detect defects. The generation of defects is detected, and the occurrence of defects is predicted and avoided in advance.
[0012]
More specifically, the solution to the above problem is realized by the following solution. The present invention, taking an illuminating means for illuminating the inspected substrate manufactured by the board manufacturing process, an imaging unit for capturing a macro image of the illuminated said inspected substrate surface by the illumination means, by the image pickup means image processing means for creating an overlay image Mareta superimposed a plurality of macro images of the inspected substrate, before Symbol superimposition statistical processing or analyzing the information of the image, the common defect of the substrate to be inspected a macro inspection apparatus characterized by comprising a process monitoring means determines abnormality detection or occurrence of abnormality prediction of the manufacturing process due to.
Further, the present invention is illuminated by the illumination means with respect to the inspected substrate manufactured in board manufacturing process, macroscopically imaged by the imaging means illuminated the inspection substrate surface by said lighting means, the imaging means in creating an image overlay with ipecac to the image processing unit each macro image of the inspected substrate of a plurality captured, the superposition image statistical processing or analysis by ipecac to the computer unit, the each It is a process monitoring method characterized by determining an abnormality of the manufacturing process due to a common defect of an inspection substrate or a prediction of occurrence of the abnormality.
Further, continuously in board manufacturing process macroscopically imaging the surface of each test board are plural manufacturing, perform acceptance judgment for each Ma Kuroga image thus shooting the image of this, the result of this determination Abnormalities in the substrate manufacturing process by creating a superimposed image by superimposing a plurality of macro images excluding NG substrates determined to be reworked or discarded, and detecting common defects appearing at the same place in the superimposed image Alternatively , a process monitoring method is characterized in that a prediction of occurrence of an abnormality is determined .
[0013]
In the present invention, by providing such a configuration, for example, individual macro images can be continuously added and displayed, and macro defects and thin film thickness variations of a substrate to be inspected can be emphasized or detected. be able to. In other words, to detect defects that cannot be detected on a single sheet basis, or to predict the occurrence of defects by statistically capturing subtle modulation of the process even when only non-defective substrates are flowing, and avoiding the occurrence of defects Is possible.
[0014]
Next, the image processing means of the present invention adds the macro images of the substrates to be inspected taken from the imaging means, and then subtracts the average density of the added images to create the superimposed image It is characterized by doing.
The image processing means of the present invention is characterized in that a corrected macro image from which background unevenness has been removed is added to the macro image of each of the substrates to be inspected taken from the imaging means.
Further, the image processing means of the present invention creates a shading image obtained by smoothing processing and shading correction for each macro image of each inspection board , and adds an image obtained by subtracting the shading image from the macro image. It is characterized by.
[0015]
Image processing means of the present invention, to keep By providing such a configuration, since is subtracted in each smoothed image when creating a superimposed image, the average density level of the superimposed image to a predetermined value Can do. Therefore, without the density value of the superimposed image is immediately saturated, it is possible to create a more effective superimposed image.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below.
FIG. 1 is a schematic configuration diagram showing an example of a macro inspection apparatus according to an embodiment of the present invention.
[0018]
In the drawing, a white light source 21 using a halogen lamp or the like as an interference light source is provided in the illumination system, and a filter rotating unit 22 is disposed in front of the white light source 21. A plurality of types of interference filters 20 are attached to the filter rotating unit 22 to enable extraction of monochromatic light suitable for observation of the substrate to be inspected.
[0019]
The light passing through the filter rotating unit 22 is given to the frost 24 through the optical fiber 23. The frost 24 converts light from the optical fiber 23 into diffused light.
[0020]
The diffused light from the frost 24 is reflected by the half mirror 25 and guided to the collimating lens 26, where it is converted into parallel light. This parallel light is uniformly irradiated onto the surface of the substrate 27 coated with a resist or the like, which is an object to be inspected.
[0021]
In this case, the light irradiated on the substrate 27 is applied to the substrate surface and reflected by the upper and lower surfaces of the thin film to form an interference image depending on the film thickness.
On the other hand, a metal halide lamp 41 and a collimating lens 42 are provided as separate illumination systems. The light emitted from the metal halide lamp 41 becomes parallel light at the collimating lens 42 and irradiates the substrate 27 at a time from obliquely above.
[0022]
In this case, the light irradiated onto the substrate 27 is diffracted by the pattern formed on the substrate surface, and forms a diffraction image depending on the pattern. Defects such as dust and scratches on the substrate can also be observed with this diffractive optical system.
[0023]
These images such as interference images, diffraction images, dust, and scratches are formed on the CCD camera 29 through the collimating lens 26 and the half mirror 25 and further through the zoom lens 28. As the imaging camera 29, a high-resolution monochrome camera is used.
[0024]
Next, the configuration of the observation system and information analysis system of the macro inspection apparatus will be described.
The CCD camera 29 is connected to the image processing device 31 through the camera controller 30.
[0025]
The image processing apparatus 31 includes a plurality of image memories M1 to Mn, performs image processing for determining pass / fail of each substrate to be inspected, and basically images of inspection substrates for 20 consecutive lots. Are superimposed to create a superimposed image (also referred to as an addition accumulated image).
[0026]
Note that the addition accumulation image is used as information for performing trend analysis of the inspection board, and an addition accumulation image for one lot or halfway to one lot and a number of addition accumulation images exceeding one lot depending on the setting. Is also an accumulated image. This addition accumulated image is created by simple addition or addition of a screen subjected to shading correction.
[0027]
In this embodiment, a method of creating a shading correction image from a raw image and adding and accumulating it in order to eliminate background unevenness of the image. Further, since the trend data is destroyed if a complete NG image is added, the image processing apparatus 31 is programmed not to add an NG image.
[0028]
In the case of simple addition, after adding the raw images obtained by the macro inspection, a certain density (average density of the added images) is subtracted to create an addition accumulated image.
[0029]
The display unit 32 includes a CRT monitor or a liquid crystal display panel connected to the image processing device 31. The display unit 32 displays the substrate to be inspected as one macro image and displays the superimposed image.
[0030]
The line monitor 35 is connected to the image processing device 31 and is a dedicated display device that displays an image when the superimposed image is always observed on the monitor. Note that a density cross-sectional image of the added accumulated image can also be displayed.
[0031]
The personal computer 33 receives information processed by the image processing device 31. In addition, overall control of the macro inspection apparatus, display / statistical analysis of inspection results, lot number of inspection board. Management / Manufacturing Equipment No. Perform recipe management.
[0032]
Further, the personal computer 33 executes statistical analysis based on information on an image of a substrate to be inspected and information on a superimposed image, so that the macro inspection apparatus functions as a process monitor. Note that other computer means such as a workstation may be used instead of the personal computer.
[0033]
The monitor 34 is a display monitor for the personal computer 33.
Next, the operation of the macro inspection apparatus according to the embodiment of the present invention configured as described above will be described.
[0034]
First, the apparatus according to the present embodiment performs pass / fail determination by performing image processing for each substrate to be inspected, as in the conventional automatic macro inspection apparatus. As a result, automatic macro inspection is executed.
[0035]
In addition, this macro inspection apparatus performs image superposition on the inspection board, displays it for use by the inspector, and automatically determines and notifies abnormality detection or prediction of occurrence of abnormality as a process monitor. Operate.
[0036]
Hereinafter, a procedure for creating an addition accumulated image will be described according to FIG.
FIG. 2 is a diagram showing a procedure for creating an addition accumulated image and detecting an abnormality or predicting an abnormality.
[0037]
First, in the image processing device 31, an image to be inspected (FIG. 2 (a1)) obtained by the macro inspection is captured and stored in the image memory M1. FIG. 2A2 shows a cross section of the xx ′ concentration at this time. Although not shown in FIG. 2 (a2), when the density exceeds a predetermined threshold, it is automatically determined and notified that an abnormality has occurred in the substrate to be inspected.
[0038]
Next, a smoothing process is performed on the image to be inspected to create a shading image (FIG. 2 (b1)) and store it in the image memory M2. An xx ′ concentration cross section is shown in FIG. The reason why the smoothing is applied here is to take out the average density level of this image.
[0039]
Next, the shading image M2 is subtracted from the inspected image M1, and is stored in the image memory M3 as a shading image (FIG. 2 (c1)). As shown in FIG. 2C2, the average level of the entire image is 0 in the xx ′ density section.
[0040]
Next, the image memory M3 is added to the addition accumulation image memory M4. The addition accumulated image (FIG. 2 (d1)) is thus obtained. An xx ′ concentration cross section is shown in FIG. The addition accumulation image memory M4 is cleared at the first lot. Furthermore, when it is desired to obtain long-term trend data over a plurality of lots, the setting is made such that data of the same board type is added as it is without being cleared.
[0041]
Until the substrate inspection of the next lot is performed, each time a new substrate is inspected, the above-described FIG. 2 (a1) to FIG. 2 (d1) are repeated, and the addition accumulation image is further accumulated.
[0042]
The above processing is performed in the image processing apparatus 31, but this added accumulated image is sent to the personal computer 33, and analysis processing is executed in the personal computer 33.
If the variation in density in each test result is random, the density only at a specific position on the added accumulated image should not be deviated from the average level, and the whole should be averaged. On the other hand, when there is an abnormality in the process or when it is not detected to be abnormal, the concentration at a specific position may always deviate from the average level when the process starts to deviate from the normal state (hereinafter also referred to as an abnormality). is there.
[0043]
Therefore, the personal computer 33 provides a threshold for density level variation, and notifies that when the added accumulated image exceeds the threshold. In this way, in the analysis processing in the personal computer 33, a warning signal is output on the monitor 34 if the variation in the density level of the added accumulated image is, for example, 30% or more of the maximum density value, and if it is 40% or more, an abnormal danger signal is output. To do. These threshold values can be changed according to the analysis conditions depending on the examination item, the type of the subject, and the like. In the determination in this analysis process, a density histogram of the accumulated image is used as will be described later.
[0044]
As a result, process monitoring is performed using trend data based on a plurality of macro images.
On the other hand, the addition accumulated image created by the image processing device 31 is displayed and output from the line monitor 35 or the display unit 32 for monitoring of the inspector.
[0045]
At the time of displaying the added accumulated image from the line monitor 35 or the display unit 32, a certain density (for example, density level 100) is added to the added accumulated image created in FIG. The data is output on the device (FIG. 2 (e1)). In some cases, the xx ′ density section of this image is displayed by the setting operation of the inspector (FIG. 2 (e2)).
[0046]
When such an added accumulated image is displayed from the line monitor 35, defects appearing at the same place in the image are displayed at a luminance level that is twice the number of added substrates. Accordingly, even a thin unevenness that cannot be recognized by a single image is gradually displayed as bright unevenness and is displayed as a defect that can be sufficiently recognized visually.
[0047]
In addition, as a macro image of a complete non-defective substrate is accumulated, the film thickness variation of the production lot becomes visible. If the variation in film thickness is uniform, there is no problem, but if it occurs in a certain place, it can be estimated that there is a problem in the manufacturing apparatus. Such a problem can be easily detected by the threshold level notification function of the personal computer 33. Further, FIG. 2 (d1), FIG. 2 (d2), FIG. 2 (e1), or FIG. 2 (e2). Specific abnormalities and signs of abnormalities can be examined in detail from the accumulated accumulated images and density cross-sectional images.
[0048]
FIG. 3 shows an example of an addition accumulated image obtained from the macro inspection apparatus according to the present embodiment. FIG. 4 shows an example of a defect map obtained from a micro inspection apparatus as a comparative example.
[0049]
From FIG. 3, it can be seen that density unevenness is caused by accumulating macro images. On the other hand, it can be seen from the comparison with FIG. 4 that the density unevenness does not necessarily correspond to micro defects. Thus, it can be seen that the present macro inspection apparatus can detect process problems that cannot be detected by the micro inspection apparatus.
[0050]
Next, detection of process abnormality using the macro inspection apparatus of the present embodiment and feedback of abnormality analysis will be described.
FIG. 5 is a diagram showing a manufacturing process of the liquid crystal panel substrate.
[0051]
The panel substrate is completed by repeating the film forming process 51, the resist coating process 52, the exposure process 53, the developing process 54, the inspection process 55 (macro inspection 55) and the etching process 56 7 to 8 times.
[0052]
In this inspection step 55, defects and abnormalities in the resist coating / exposure / development steps 52, 53 and 54 are inspected. The substrate determined to be NG here is restarted from the resist coating step 52 (referred to as rework) or discarded after the resist is peeled off.
[0053]
The macro inspection apparatus captures defects generated in the photolithography process such as resist application failure, exposure failure, development failure, dust, and transport defect. The detected defect contents are analyzed and classified, and finally, the inspector finds out the cause of the defect and feeds it back to the manufacturing apparatus in each of the resist coating / exposure / development steps 52, 53, and 54. Become.
[0054]
In this process, in the present inspection apparatus, the addition accumulated image of the inspected substrate obtained by the macro inspection is also displayed on the line monitor 35. Therefore, the inspector can visually check the determination information based on the added accumulated image and monitor the process.
[0055]
Further, as described above, an automatic caution warning notification is also made by the personal computer 33.
FIG. 6 is a diagram showing a density histogram of the added accumulated image.
[0056]
That is, if the automatic analysis function for the added accumulated image of the personal computer 33 takes the density histogram of the superimposed image shown in FIG. 6, and the difference between the maximum density and the minimum density is equal to or greater than a set value (for example, 40%), It is determined that the variation in film thickness is too large, and a caution signal is output to an inspector or a manufacturing apparatus as a forefront of the occurrence of a defect. The inspector performs process monitoring with reference to the notification information of the warning alarm.
[0057]
As described above, in the macro inspection apparatus and the process monitoring method according to the embodiment of the present invention, the accumulated image is created by superimposing the inspection images of the substrate to be inspected, and this is statistically processed or analyzed to monitor the process. Therefore, the common defect appearing on the lot substrate can be detected.
[0058]
In addition, since the occurrence location of the common defect can be specified, the cause of the occurrence of the defect can be easily traced. Furthermore, defects such as thin unevenness that cannot be recognized by a single substrate can also be detected.
[0059]
Regardless of the quality of each substrate, the macro image is added and accumulated and output to the monitor, so it is possible to grasp the sign of defect occurrence from the degree of variation such as film thickness unevenness, and to predict the occurrence of defects. it can.
[0060]
In this way, in the photolithography process for wafers or liquid crystal glass substrates, etc., the production apparatus can be managed so as to predict the occurrence of defects and stop the occurrence in advance, so that manufacturing costs can be reduced and production efficiency can be improved. .
[0061]
In addition, this invention is not limited to said each embodiment, It can change variously in the range which does not deviate from the summary.
In the present embodiment, the image of the inspection substrate for 20 consecutive lots is superimposed to create a superimposed image. However, the number of sheets to be superimposed is not limited to this number, and the number of overlays can be changed as appropriate. Can be adjusted.
[0062]
Furthermore, the present invention can be applied to inspection of various substrates such as substrates for flat panels such as liquid crystal panels, PDPs, and FEDs, and substrates such as wafers for semiconductor elements.
[0063]
Further, the method described in the embodiment is stored in a storage medium such as a magnetic disk (floppy disk, hard disk, etc.), an optical disk (CD-ROM, DVD, etc.), a semiconductor memory, etc. as a program that can be executed by a computer. In addition, it can be transmitted and distributed by a communication medium. A computer that implements this apparatus reads a program recorded in a storage medium and executes the above-described processing by controlling the operation by this program.
[0064]
【The invention's effect】
As described above in detail, according to the present invention, since the addition accumulation image of the macro image is created and analyzed, it is possible to detect a defect that cannot be detected in a single unit, or even when only a non-defective substrate is flowing, It is possible to provide a macro inspection apparatus and a process monitoring method capable of statistically capturing subtle modulation of a process, predicting the occurrence of a defect in advance, and avoiding the occurrence of the defect.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram showing an example of a macro inspection apparatus according to an embodiment of the present invention.
FIG. 2 is a diagram showing a procedure for creating an addition accumulated image and detecting an abnormality or predicting an abnormality.
FIG. 3 is a view showing an example of an addition accumulated image obtained from the macro inspection apparatus according to the embodiment;
FIG. 4 is a comparison diagram showing an example of a defect map obtained from a micro inspection apparatus.
FIG. 5 is a view showing a manufacturing process of a liquid crystal panel substrate.
FIG. 6 is a diagram showing a density histogram of an addition accumulated image.
[Explanation of symbols]
21 ... White light source 22 ... Filter rotating unit 23 ... Optical fiber 24 ... Frost 25 ... Half mirror 26 ... Collimating lens 27 ... Substrate 28 ... Zoom lens 29 ... CCD camera 30 ... Camera controller 31 ... Image processing device 32 ... Display unit 33 ... PC 34 ... Monitor 35 ... Line monitor 41 ... Metal halide lamp 42 ... Collimating lens

Claims (10)

Illuminating means for illuminating the inspected substrate manufactured by the board manufacturing process,
Imaging means for capturing a macro image of the surface of the substrate to be inspected illuminated by the illumination means;
Image processing means for creating a superimposed image obtained by superposing macro images of the plurality of substrates to be inspected taken by the imaging means;
Statistical processing or analyzing the information before Symbol superimposed image, the process monitoring means for determining the prediction of abnormality detection or abnormality of the manufacturing process due to the common defect of the substrate to be inspected,
A macro inspection apparatus comprising: The process monitoring unit receives the superimposed image processed by said image processing means in the computer means, the concentration of the specific position of the superimposition analyzing processing density information of the image, the superimposed in the image in the computer unit 2. The macro inspection apparatus according to claim 1, wherein the information and the threshold value are compared to automatically determine abnormality detection or prediction of occurrence of abnormality in the manufacturing process . Claim wherein the process monitoring unit, when the difference between the maximum and minimum densities to the concentration histograms of the superimposed image is larger than a predetermined value, characterized in that for outputting a caution signal to the manufacturing apparatus predicts occurrence of defects The macro inspection apparatus according to either 1 or 2 . The image processing means makes a pass / fail determination for each of the substrates to be inspected taken by the imaging means, and the process monitoring means excludes an NG board that is determined to be rework or discarded by the pass / fail determination. The macro inspection apparatus according to claim 1, wherein only the two are superposed. The image processing means creates a superimposed image by adding a plurality of defects that cannot be detected by a single macro image of the substrate to be inspected , and the process monitoring means is provided at the same location in the superimposed image. 3. The macro inspection according to claim 1, wherein the occurrence of a defect in the manufacturing process is detected or the occurrence of the defect is predicted from a luminance level that is obtained by adding the number of appearing defects to the number of substrates. apparatus. The image processing means adds the macro images of the respective substrates to be inspected taken from the imaging means, and then subtracts the average density of the added images to create the superimposed image. The macro inspection apparatus according to claim 1. 2. The macro inspection apparatus according to claim 1, wherein the image processing unit adds a corrected macro image from which background unevenness has been removed to a macro image of each of the substrates to be inspected taken from the imaging unit. apparatus. Wherein the image processing means, wherein with respect to the respective macro image of each test board, to create a shading image smoothing process and the shading correction, characterized by adding an image obtained by subtracting the shading image from the macro image The macro inspection apparatus according to claim 1. Illuminated by the illumination means with respect to the inspected substrate manufactured in board manufacturing process,
The test substrate surface illuminated by said illuminating means macroscopically imaged by the imaging means, the image superimposed with ipecac a plurality of the macro image of the substrate to be inspected picked up by the image pickup means to the image processing unit Create
A process monitoring method characterized in that the superimposed image is taken into a computer means and subjected to statistical processing or analysis to determine an abnormality in the manufacturing process or a prediction of occurrence of the abnormality caused by a common defect of each substrate to be inspected . Successively with board manufacturing process macroscopically imaging the surface of each test board are plural manufacturing, perform acceptance judgment for each Ma Kuroga image thus shooting the image of this, rework or by the determination result Abnormality or abnormality of the substrate manufacturing process by creating a superimposed image by superimposing a plurality of the macro images excluding the NG substrate determined to be discarded, and detecting a common defect appearing at the same location of the superimposed image A process monitoring method characterized by determining occurrence prediction .

Images