WO2014208362A1

WO2014208362A1 - Applicator device and height detection method

Info

Publication number: WO2014208362A1
Application number: PCT/JP2014/065711
Authority: WO
Inventors: 博明大庭
Original assignee: Ntn株式会社
Priority date: 2013-06-25
Filing date: 2014-06-13
Publication date: 2014-12-31
Also published as: JP2015007564A; JP6189102B2; CN105247318A; CN105247318B

Abstract

A controlling computer (11) of this defect repair device (1) (applicator device) positions an objective lens (16) above an ink application part comprising ink applied to the surface of a substrate (7), then captures an image while moving a Z stage (8), calculates, for each of a plurality of pixels constituting the captured image, a Z stage position (focal location of the pixel) at which a contrast value C reaches a peak, and on the basis of the calculated Z stage position, calculates the height of the ink application part. Thus, the height of the ink application part can be detected quantitatively.

Description

Coating apparatus and height detection method

The present invention relates to a coating apparatus and a height detection method, and in particular, a coating apparatus that applies a liquid material to the surface of a substrate, and a height detection method that detects the height of a coating portion made of the liquid material applied to the surface of the substrate. About.

In recent years, the number of pixels has increased with the increase in size and definition of liquid crystal displays, making it difficult to produce liquid crystal displays without defects, and the probability of occurrence of defects has also increased. Under such circumstances, in order to improve the yield, a defect correcting apparatus that corrects a defect that occurs in the manufacturing process of the liquid crystal color filter substrate has become indispensable for the production line.

23 (a) to 23 (c) are diagrams showing defects that occur in the manufacturing process of the liquid crystal color filter substrate. 23A to 23C, the liquid crystal color filter substrate includes a transparent substrate, a lattice pattern called a black matrix 51 formed on the surface thereof, and a plurality of sets of R (red) pixels 52, G ( Green) pixel 53 and B (blue) pixel 54. In the manufacturing process of the liquid crystal color filter substrate, as shown in FIG. 23A, the white defect 55 in which the color of the pixel or the black matrix 51 is lost, or the adjacent pixel and the color as shown in FIG. Are mixed, or a black defect 56 in which the black matrix 51 protrudes from the pixel, or a foreign object defect 57 in which a foreign object adheres to the pixel as shown in FIG.

As a method of correcting the white defect 55, an ink having the same color as that of the pixel in which the white defect 55 exists is attached to the tip of the application needle by the ink application mechanism, and the ink attached to the tip of the application needle is applied to the white defect 55 There is a method to correct by applying. Further, as a method of correcting the black defect 56 and the foreign object defect 57, after the defective portion is laser-cut to form a rectangular white defect 55, the ink applied to the tip of the application needle is removed by the ink application mechanism. There is a method in which the defect 55 is applied and corrected (see, for example, Patent Document 1 (Japanese Patent Laid-Open No. 2009-122259)).

Also, there is a method of capturing an image of a region including a defect before and after the correction process, comparing the brightness of the image before and after the correction process, and detecting an abnormality in the correction process based on the comparison result (for example, Patent Document 2 No. 2009-237086)).

JP 2009-122259 A JP 2009-237086 A

A liquid crystal display is obtained by bonding a TFT substrate on which an electronic circuit is formed and a liquid crystal color filter substrate that expresses the color of a pixel, and sealing liquid crystal between two substrates. However, if protrusions higher than a predetermined height are present on the surface of at least one of the two substrates, the liquid crystal cannot be normally sealed between the two substrates. For this reason, before bonding the two substrates, it is necessary to determine whether the bonding is possible by inspecting the presence or absence of protrusions on the surface of the substrate.

When the white defect 55 is corrected by applying high viscosity ink, the ink application portion made of the applied ink becomes a protrusion on the surface of the liquid crystal color filter substrate. Therefore, it is necessary to detect the height of the ink application part after correcting the white defect 55. However, the method of Patent Document 2 cannot quantitatively detect the height of the ink application portion even if a planar defect such as the color and density of the ink application portion or the size and shape can be detected. It was.

Therefore, a main object of the present invention is to provide a coating apparatus capable of quantitatively detecting the height of the coating part and a height detection method.

The coating apparatus according to the present invention is a coating apparatus that applies a liquid material to the surface of a substrate, and includes an observation optical system that observes the surface of the substrate via an objective lens, and an image of the surface of the substrate via the observation optical system. A head unit including an image pickup apparatus for picking up images, a coating mechanism for applying a liquid material to the surface of the substrate, and the head unit and the substrate are relatively moved to position the head unit at a desired position above the surface of the substrate. The positioning device, the positioning device, and the imaging device are controlled to position the objective lens above the application portion made of a liquid material applied to the surface of the substrate, and then the application portion and the objective lens are moved relative to each other in the vertical direction. A height detection unit is provided that captures an image, obtains a focal position for each of a plurality of pixels constituting the captured image, and obtains a height of the application unit based on the obtained focal position.

Further, the height detection method according to the present invention includes an observation optical system for observing the surface of the substrate through the objective lens, an imaging device for capturing an image of the surface of the substrate through the observation optical system, and a surface of the substrate. In a coating apparatus comprising: a head unit including a coating mechanism that coats a liquid material; and a positioning device that relatively moves the head unit and the substrate to position the head unit at a desired position above the surface of the substrate. A height detection method for detecting the height of an application part made of a liquid material applied to the surface of the apparatus, wherein the positioning device and the imaging device are controlled to position the objective lens above the application part, and An image is picked up while relatively moving the objective lens in the vertical direction, a focal position is obtained for each of a plurality of pixels constituting the picked-up image, and the height of the coating part is obtained based on the obtained focal position. .

In the coating apparatus and the height detection method according to the present invention, an image is captured while the coating unit and the objective lens are relatively moved in the vertical direction, and a focal position is obtained and obtained for each of a plurality of pixels constituting the captured image. The height of the application part is obtained based on the focal position. Therefore, the height of the application part can be detected quantitatively.

It is a perspective view which shows the whole structure of the defect correction apparatus by Embodiment 1 of this invention. It is a perspective view which shows the structure of the ink application | coating mechanism shown in FIG. It is a figure which shows operation | movement of the ink application | coating mechanism shown in FIG. It is a figure which shows the surface of the liquid crystal color filter substrate shown in FIG. It is a figure which shows the defect detection operation | movement of the control computer shown in FIG. It is a figure which shows the calculation method of the contrast value of the image by the computer for control shown in FIG. It is a figure which shows the pixel height detection method by the computer for control shown in FIG. It is a figure which shows the test conditions of the ink application part apply | coated by the ink application mechanism shown in FIG. It is a perspective view which shows the principal part of the defect correction apparatus by Embodiment 2 of this invention. It is a figure which shows the structure of the ink application | coating mechanism shown in FIG. It is a figure which shows the principal part of the defect correction apparatus by Embodiment 3 of this invention. It is a figure which shows the pixel height detection method using the Mirau type | mold interference objective lens shown in FIG. FIG. 11 is a diagram for explaining a problem of the third embodiment. FIG. 10 is another diagram for explaining the problem of the third embodiment. It is a figure for demonstrating the principle of the height detection method by Embodiment 4 of this invention. It is another figure for demonstrating the principle of the height detection method by Embodiment 4. FIG. It is a figure which shows the peak position of a certain line on an image. FIG. 18 is a diagram showing a zero point of a phase δ nearest to the peak position shown in FIG. 17. It is a figure which shows the deviation | shift amount of the peak position shown in FIG. 16, and the 0 point shown in FIG. It is a figure which shows the deviation | shift amount after correction. It is a figure which shows 0 point after correction. It is a figure which shows the inspection item of an applicator needle. It is a figure which shows the defect of a liquid crystal color filter.

[Embodiment 1]
[Device configuration]
As shown in FIG. 1, the defect correcting apparatus 1 according to the first embodiment of the present invention includes an observation optical system 2, a CCD camera 3, a cutting laser device 4, an ink application mechanism 5, and an ink curing light source 6. A correction head portion, a Z stage 8 that moves the correction head portion in a direction perpendicular to the liquid crystal color filter substrate 7 to be corrected (Z-axis direction), and a Z stage 8 that is mounted and moved in the X-axis direction. An X stage 9; a Y stage 10 on which the substrate 7 is mounted and moved in the Y-axis direction; a control computer 11 for controlling the operation of the entire apparatus; a monitor 12 for displaying an image taken by the CCD camera 3; And an operation panel 13 for inputting a command from an operator to the control computer 11.

The observation optical system 2 includes a light source for illumination, and observes the surface state of the substrate 7 and the state of the correction ink applied by the ink application mechanism 5. An image observed by the observation optical system 2 is converted into an electrical signal by the CCD camera 3 and displayed on the monitor 12. The cutting laser device 4 removes unnecessary portions on the substrate 7 by irradiating them with laser light via the observation optical system 2.

The ink application mechanism 5 corrects the white defect generated on the substrate 7 by applying correction ink. The ink curing light source 6 includes, for example, a CO ₂ laser, and cures the correction ink applied by the ink application mechanism 5 by irradiating it with laser light.

This apparatus configuration is an example. For example, the Z stage 8 on which the observation optical system 2 or the like is mounted is mounted on the X stage, the X stage is mounted on the Y stage, and the Z stage 8 can be moved in the XY directions. A configuration called a gantry system may be used, and any configuration may be used as long as the Z stage 8 on which the observation optical system 2 and the like are mounted can be moved relative to the correction target substrate 7 in the XY directions.

Next, an example of an ink application mechanism using a plurality of application needles will be described. FIG. 2 is a perspective view showing the main parts of the observation optical system 2 and the ink application mechanism 5. In FIG. 2, the defect correcting apparatus 1 includes a movable plate 15, a plurality of (for example, five) objective lenses 16 having different magnifications, and a plurality of (for example, five) application units 17 for applying different color inks. With.

The movable plate 15 is provided between the lower end of the observation barrel 2a of the observation optical system 2 and the substrate 7 so as to be movable in the X-axis direction and the Y-axis direction. Further, the movable plate 15 is formed with five through holes 15a corresponding to the five objective lenses 16, respectively.

The five through holes 15a are arranged at predetermined intervals in the Y-axis direction. Each objective lens 16 is fixed to the lower surface of the movable plate 15 so that its optical axis coincides with the center line of the corresponding through hole 15a. The optical axis of the observation barrel 2a and the optical axis of each objective lens 16 are arranged in the Z-axis direction perpendicular to the X-axis direction and the Y-axis direction. By moving the movable plate 15, the objective lens 16 having a desired magnification can be disposed below the observation barrel 2a.

The five coating units 17 are fixed to the lower surface of the movable plate 15 at a predetermined interval in the Y-axis direction. Each of the five coating units 17 is disposed adjacent to the five objective lenses 16. By moving the movable plate 15, it is possible to arrange the desired coating unit 17 above the white defect to be corrected.

3 (a) to 3 (c) are views showing the main part from the direction A in FIG. 2, and showing the ink application operation. The application unit 17 includes an application needle 18 and an ink tank 19. First, as shown in FIG. 3A, the application needle 18 of the desired application unit 17 is positioned above the white defect to be corrected. At this time, the tip of the application needle 18 is immersed in the correction ink in the ink tank 19.

Next, as shown in FIG. 3B, the application needle 18 is lowered and the tip of the application needle 18 protrudes from the hole at the bottom of the ink tank 19. At this time, correction ink is attached to the tip of the application needle 18. Next, as shown in FIG. 3C, the application needle 18 and the ink tank 19 are lowered to bring the tip of the application needle 18 into contact with the white defect, and the correction ink is applied to the white defect. Thereafter, the state returns to the state of FIG.

The ink application mechanism using a plurality of application needles is not described in detail since various other techniques are known. For example, it is shown in Patent Document 1 (Japanese Patent Laid-Open No. 2009-122259). The defect correction apparatus 1 can correct a defect using ink of a desired color among a plurality of inks by using, for example, a mechanism as shown in FIG. 2 as the ink application mechanism 5. The defect can be corrected using an application needle having a desired application diameter among the application needles.

[Defect detection process]
FIG. 4 is a view showing the surface of the liquid crystal color filter substrate 7. In FIG. 4, the liquid crystal color filter substrate 7 includes a plurality of picture elements PC formed on the surface of a glass substrate. The beginning DS of the picture element PC and the end DE of the picture element PC exist at the intersection positions of the black matrix portions BM formed vertically and horizontally. The start DS of the picture element PC is referred to as a color filter position. The control computer 11 specifies the position of this color filter. In addition, the range from the start DS of the picture element PC to the end DE of the picture element surrounded by a square in FIG.

In the binary image, a set of pixels having a value of 1 in the picture element PC is a color filter portion (shown by a color filter portion CF in the figure), and a set of pixels having a value of 0 (hatched portion in the figure) is This is a black matrix portion (indicated by a black matrix portion BM in the figure) of the picture element PC. Each picture element PC has one of RGB (Red, Green, Blue) different from each other, and is repeatedly formed at a constant cycle.

FIGS. 5A and 5B are diagrams showing operations when the control computer 11 detects a defect in the horizontal direction of the input image. The control computer 11 detects a defective portion based on the brightness of the pixels of the color filter. More specifically, the control computer 11 sets the brightness f (x, y) at the position (x, y) in the input image, where P is the interval between picture elements arranged periodically, that is, at equal intervals. On the other hand, a comparative inspection is performed as shown by the following formula (1).

As described above, the control computer 11 compares the luminance f (x, y) with the luminance f (x−P, y) before one cycle and the luminance f (x + P, y) after one cycle. Here, s-p (x, y) is a comparison result between f (x, y) and f (x-P, y), and s + p (x, y) is f (x, y) and f. The comparison result with (x + P, y) is shown.

The control computer 11 compares sH (x, y) with the slice level Td when the signs of sp (x, y) and s + p (x, y) match. Further, when the signs of sp (x, y) and s + p (x, y) do not match, the control computer 11 determines whether the position (x−P, y) or the position (x + P, y). The position (x, y) is excluded from the inspection target because it is highly likely to be erroneously detected as a pixel defect in FIG. With such a configuration, an error in defect detection due to noise in the input image can be prevented.

Then, if sH (x, y) is equal to or greater than Td, the control computer 11 determines that the pixel at the position (x, y) is a defect and stores the result in dH (x, y). In dH (x, y), a pixel having a value of 1 indicates a defect, and a pixel having a value of 0 indicates normal.

Next, the control computer 11 calculates the centroid position of the portion having a value of 1 (ie, white defect), and the X stage 9 and the Y stage so that the coordinates of the calculated centroid position coincide with the center of the screen of the monitor 12. 10 is controlled. Further, the control computer 11 determines the color of the ink to be applied to the white defect. Further, the control computer 11 calculates the ink application position in the white defect. Such a defect detection step is disclosed in, for example, Japanese Patent Application Laid-Open No. 2007-233299.

Thereafter, the control computer 11 selects the application unit 17 for applying the ink of the determined color, contacts the tip of the application needle 18 of the application unit 17 with the calculated ink application position, and determines the determined color. Apply correction ink to white defects. The correction ink applied to the white defect is cured by irradiating the light of the ink curing light source 6, and the correction of the white defect is completed.

[Height detection process]
In this step, the control computer 11 controls the defect correcting device 1 to obtain the height of the ink application portion made of the correction ink applied to the white defect and cured. The height detection method of the first embodiment is suitable for detecting the height of the ink application part that is higher than the focal depth of the objective lens 16.

In this height detection method, utilizing the fact that the contrast of the image is maximized at the focal position, the Z stage 8 is moved relative to the ink application part, and the contrast is maximized for each pixel of the image. The stage position is obtained and the position is used as the height information of the pixel.

First, the search procedure is shown. The Z stage 8 is moved to the search start position. If the current position is Zp and the search range is Δ, for example, it moves to Zp−Δ / 2. Here, the minus direction of the Z stage 8 is the direction approaching the substrate 7, and the search is performed in the plus direction from the initial position, that is, the direction away from the substrate 7. Therefore, a range of Δ is searched in the positive direction from the initial position Zp−Δ / 2. Note that the search direction is not necessarily a direction away from the substrate 7, and may be a direction approaching.

The control computer 11 starts sampling the image after the Z stage 8 starts to move and reaches a constant speed state. Sampling is performed at regular intervals. Preferably, sampling can be performed more accurately by performing the period of the vertical synchronizing signal of the CCD camera 3. The Z stage 8 moves at a predetermined speed v (μm / second). The velocity v satisfies the condition of D ≦ (1 / F) × v, where D (μm) is the depth of focus of the objective lens 16 to be used, and F (Hz) is the frequency of the vertical synchronization signal of the CCD camera 3. It is desirable. This is because the depth of focus is the length of the region that appears to be in focus, so that the image cannot be changed unless it moves at least D (μm) during the sampling period.

As described above, the image is sampled while moving the Z stage 8 within the search range, and the contrast value C of the image is calculated for each pixel of the acquired image. As shown in FIG. 6, the contrast value C is the luminance fi (x + a) of the pixel (x + a, y + b) separated by (a, b) vertically and horizontally with respect to the luminance fi (x, y) of the target pixel (x, y). , Y + b) and dx _xy , dy _xy , the following formula (2) is calculated.

In this equation (2), (H, W) indicates the number of pixels in the horizontal and vertical directions of the image. Further, fi (x, y) indicates the luminance of the pixel of the i-th sampled image, i is an image number given in the order of acquisition, and takes values of i = 1, 2,. .

7A is a diagram showing the relationship between the Z stage position and the contrast value C, and FIG. 7B is a diagram showing the relationship between the Z stage position and its speed. The contrast value C has a mountain shape as shown in FIG. 7A, and the peak of the mountain is the focal position. Even when the Prewitt operator and Sobel operator generally used in image processing are applied to an image and the average luminance value of the image after application is plotted, the same tendency as in FIG. In other words, any image feature that exhibits a tendency similar to that of Equation (2) may be used. Since the image is sampled at least every D (μm), the true focal position is likely to exist between samples. For this reason, interpolation is performed using data in the vicinity of the position where the contrast value C is maximum, and an accurate focal position is obtained by approximation.

The data in the vicinity of the focal position can be approximated by a quadratic function or a Gaussian function because it shows a symmetrical mountain-shaped tendency centered on the focal position. Function approximation is performed by the Newton method or the like using the Z stage coordinates near the focal position and the contrast value, and the peak position is interpolated from the obtained function to obtain the height of the corresponding pixel. In addition to function approximation, the center of gravity position may be obtained using the contrast value around the peak, and the obtained center of gravity position may be set as the height of the corresponding pixel.

[Height inspection process]
In this step, the ink application part is extracted based on the images before and after application, and the heights of the extracted ink application part and the reference part are compared. For example, as described in Patent Document 2 (Japanese Patent Laid-Open No. 2009-237086), the brightness of the image before and after application is compared, and the ink application part is extracted based on the comparison result. Let the extraction result of the ink application part be b (x, y). b (x, y) is a function that returns 1 if the pixel at the position (x, y) is an ink application portion, and returns 0 otherwise.

The reference portion is a normal portion of the substrate 7 where the correction ink is not applied, and is extracted from an image before or after application. The center coordinates (Δx, Δy) of the reference portion with respect to the application start point and the vertical and horizontal sizes (w, h) are determined in advance. Here, an image in which the height information obtained in the height detection step is stored is h (x, y), the coordinates of the application start point are (xs, ys), and the height of the reference portion is (xs + Δx, The average value of the heights in the range of (± w / 2, ± h / 2) with ys + Δy) as the center. Note that the reference portion is not limited to the above method, and for example, a characteristic portion of the substrate 7 may be detected by pattern matching or the like, or may be set in an area offset from a detection position obtained by pattern matching. .

The average height of the reference portion obtained as described above is set as h0. H0 is subtracted from the height image h (x, y), and the subtraction result is set as h '(x, y). Subsequently, the total value, the maximum value, the minimum value, the variance value, and the average value of h ′ (x, y) of the pixel indicating the value 1 of the ink application part b (x, y) extracted earlier are calculated. Note that the vertical and horizontal dimensions of one pixel are (mx, my). The unit is nm.

The total value corresponds to the volume of the ink application part, and is effective for checking whether a predetermined ink application amount can be secured or whether the upper limit is exceeded. The total value is calculated using the following formula.

The maximum value is the maximum value of h ′ (x, y) of a pixel having a b (x, y) value of 1, and is effective for checking whether the height of the ink application part exceeds the upper limit. It is.

The minimum value is the minimum value of h ′ (x, y) of a pixel having a b (x, y) value of 1, and is effective for checking whether or not a certain thickness can be secured.

The dispersion value is effective when evaluating the uniformity of the height of the ink application part. The variance value is calculated according to the following equation (4).

The average value is effective for checking whether or not a certain height is secured over the entire ink application part. The average value is calculated according to the following equation (5).

The control computer 11 determines whether or not the ink application unit is normal based on at least one of the calculated total value, maximum value, minimum value, variance value, and average value.

The defect correction apparatus 1 according to the first embodiment has a function of registering inspection items in advance in the order of application, and the inspection items and the allowable range can be changed depending on the type of application needle, the substrate 7, and the correction ink. Yes.

FIG. 8 is a diagram showing inspection conditions when ink is applied by the ink application mechanism 5 shown in FIG. The ink application mechanism 5 has five application needles, and inspection conditions can be registered for each application needle. The registered content is referred to when application is performed with the corresponding application needle. AND is passed when all the specified conditions are met, and OR is passed when any one of the conditions is met.

適用 Applicable when numerical values are specified in the columns of “Total Value”, “Maximum Value”, “Minimum Value”, “Dispersion Value”, and “Average Value”. In this example, two types of settings “AND” or “OR” are possible for “final determination”. The numeric field is specified as a (lower limit, upper limit) pair. When both the lower limit value and the upper limit value are numerically designated, the condition is satisfied when the value of the corresponding inspection item is not less than the lower limit value and less than the upper limit value. When the lower limit value is “−”, the condition is satisfied when the value is less than or equal to the upper limit value. When the upper limit is “−”, the condition is satisfied when the value is equal to or greater than the lower limit. It is not judged when both are blank.

In the first embodiment, an image is captured while the ink application unit and the objective lens 16 are relatively moved in the vertical direction, a focal position is obtained for each of a plurality of pixels constituting the captured image, and the obtained focal position is obtained. Based on this, the height of the ink application part is obtained. Therefore, the height of the ink application part can be easily and accurately detected quantitatively. As a result, accurate inspections such as changes in the viscosity of the corrected ink and detection of an abnormal state of the ink application mechanism 5 can be performed, which can contribute to an improvement in manufacturing process yield.

In the first embodiment, the case where the present invention is applied to the detection of the height of the ink application portion made of the correction ink applied to the liquid crystal color filter substrate 7 has been described, but the present invention is not limited to this. It goes without saying that the present invention can be applied to the detection of the height of an application part made of a liquid material applied to a substrate. For example, the present invention can be applied to the detection of the height of a paste application portion made of a conductive paste applied to a disconnection defect portion of a wiring on the surface of a substrate such as a TFT substrate or a printed circuit board.

[Embodiment 2]
FIG. 9 is a diagram showing a main part of the defect correcting apparatus according to the second embodiment of the present invention, and is a diagram contrasted with FIG. Referring to FIG. 9, this defect correcting device is different from defect correcting device 1 of the first embodiment in that coating unit 17 is replaced with electrostatic ink jet device 20. The electrostatic inkjet device 20 is fixed to the lower surface of the movable plate 15.

FIG. 10 is a diagram showing a main part of the electrostatic ink jet apparatus 20. In FIG. 10, the electrostatic inkjet device 20 includes an inkjet nozzle 21, a pulse voltage generator 22, and a controller 23. The nozzle 21 is a glass tube that is stretched to have a very small tip diameter. Conductive correction ink 24 is injected into the nozzle 21, and a pulse voltage VP output from the pulse voltage generator 22 can be applied to the correction ink 24. The substrate 7 is fixed horizontally on the Y stage 10. A desired target position on the surface of the substrate 7 can be positioned below the nozzle 21 by driving the stages 8 to 10.

During the drawing operation, the tip 21a of the nozzle 21 and the surface of the substrate 7 face each other with a minute drawing distance d. When the pulse voltage VP is applied to the correction ink 24 injected into the nozzle 21 in this state, a conical tailor cone 24a is formed from the tip 21a of the nozzle 21 toward the substrate 7, and the surface of the substrate 7 is formed from the top of the tailor cone 24a. A jet flow (liquid column) 24b that reaches 1 is generated, and a part of the correction ink 24 moves onto the surface of the substrate 7 to form droplets 24c. By moving the substrate 7 by the X stage 9 and the Y stage 10, an ink application portion having a desired shape can be formed on the surface of the substrate 7. Since other configurations and operations are the same as those in the first embodiment, description thereof will not be repeated. Also in this second embodiment, the same effect as in the first embodiment can be obtained.

In addition, as another application mechanism, there is a dispenser (not shown). Which coating mechanism is used may be appropriately selected according to the object and the liquid material.

[Embodiment 3]
In the defect correction apparatus according to the third embodiment, a height detection method different from the height detection method described in the first embodiment is employed. In this height detection method, a two-beam interference objective lens is used instead of the objective lens 16, and the interference fringe intensity is maximized at the focal position, and the Z stage 8 is moved relative to the substrate 7. Then, an interference fringe image is picked up, a Z stage position where the interference intensity is maximized is obtained for each pixel, and the position is set as the height of the pixel. This height detection method is suitable for detecting a minute height of several μm or less.

The two-beam interference objective lens separates the white light emitted from the light source into two light beams and irradiates one on the surface of the object and the other on the reference surface to interfere the reflected light from both surfaces. is there. In the third embodiment, a Mirau interference objective lens is used, but a Michelson type interference linique type interference objective lens may be used.

Also, a white light source is used as the light source. This is because the brightness of the interference fringes is maximized only at the focal position of the lens unlike a single wavelength light source such as a laser, and is therefore suitable for measuring the height.

FIG. 11 shows a layout of the optical elements of the observation optical system 2 when the Mirau-type interference objective lens 30 is used. The Mirau interference objective lens 30 includes a lens 31, a reference mirror 32, and a beam splitter 33. At the same time as switching the objective lens 16 to the Mirau-type interference objective lens 30, a filter 36 is inserted by the filter switching device 35 into the exit portion of the falling light source 34. When light passes through the filter 36, white light having a center wavelength λ (nm) is obtained.

The light that has passed through the filter 36 is reflected by the half mirror 37 toward the lens 31. The light incident on the lens 31 is divided by the beam splitter 33 into light that passes in the direction of the substrate 7 and two light that reflects in the direction of the reference mirror 32. The light reflected by the surface of the substrate 7 and the reference mirror 32 is again merged by the beam splitter 33 and condensed by the lens 31. Thereafter, the light emitted from the lens 31 passes through the half mirror 37 and then enters the imaging surface 3 a of the CCD camera 3 through the imaging lens 38.

Normally, the Mirau-type interference objective lens 30 is moved in the optical axis direction by the Z stage 8 to cause an optical path length difference between the surface reflected light of the substrate 7 and the surface reflected light of the reference mirror 32. The CCD camera 3 captures an interference fringe generated by the optical path length difference while moving the Mirau interference objective lens 30. The intensity, that is, the brightness of the interference fringes is maximized when the reflected light from the substrate 7 and the reflected light from the reference mirror 32 are equal in length. At this time, the surface of the substrate 7 is in focus.

In addition to the Z stage 8, the substrate 7 itself is moved up and down on the table, or a piezo table or the like is attached to the connecting portion between the Mirau interference objective lens 30 and the observation optical system 2 to set the upper and lower positions of the Mirau interference objective lens 30. You may adjust.

As in the first embodiment, the search range is Δ, and the search is performed in the positive direction from the initial position Zp−Δ / 2, that is, in the direction in which the Z stage 8 moves away from the substrate 7. As in the first embodiment, the search direction does not necessarily need to be a direction away from the substrate 7 and may be a direction approaching.

As in the first embodiment, the sampling of the image is started after the Z stage 8 starts to move and reaches a constant speed state. Further, if sampling is performed at the period of the vertical synchronization signal of the CCD camera 3, an interference fringe image can be sampled more accurately.

The Z stage 8 moves at a predetermined speed v (μm / second), and the moving speed is determined as follows. Assuming that the center wavelength of white light is λ (μm) and the frequency of the vertical synchronizing signal of the CCD camera 3 is F (Hz), the velocity v is (λ /) during the sampling period (1 / F) second of the image. 8) Set the speed to move by μm. That is, v = (λ / 8) × F. This speed v corresponds to π / 2 in white light phase increment. It is known that the peak point of the interference fringe intensity can be easily detected by changing the phase by π / 2.

When the image is sampled while changing the phase by π / 2, the contrast value Mi of the interference fringe intensity is calculated using the following equation (6) using five images.

Here, fi (x, y) indicates the value of the pixel at the position (x, y) of the image fi. I is a number assigned to an image in the order of acquisition, and takes a value of i = 1, 2,.

12A is a diagram showing the relationship between the image number i and the pixel value fi (x, y), and FIG. 12B is a diagram showing the relationship between the pixel number i and the contrast value Mi. c) is a diagram showing the relationship between the position of the Z stage 8 and the speed. 12 (a) to 12 (c), fi (x, y) and Mi show peaks in the vicinity of the image p. This peak point is the focal position of the pixel (x, y). Since Mi shows a symmetrical mountain-shaped tendency around the peak point, a curve representing Mi can be approximated by a quadratic function or a Gaussian function as in the first embodiment.

Suppose that an image in which the maximum value of Mi is stored is Mmax (x, y), and an image in which the image number indicating the maximum value is stored is I (x, y). Before starting the measurement, all pixels of Mmax (x, y) are set to 0. Further, −1 is set to all the pixels of I (x, y). During measurement, every time Mi (x, y) is calculated, Mi (x, y) is compared with Mmax (x, y). If Mi (x, y) is larger, then Mi is set to Mmax (x, y). Set (x, y) to i (x, y). When acquisition of all images within the search range is completed, an image number in the vicinity of the peak point of each pixel is stored in I (x, y).

Finally, an accurate peak point is obtained by function approximation using a total of (2n + 1) images of ± n sheets around the image p near the peak point. Let j be the number of (2n + 1) images. Since the interference fringe amplitude value Mj (x, y) of each image is obtained during the measurement, the second order is obtained by the Newton method or the like using (2n + 1) amplitude values Mj (x, y) and the image number j. Approximate with a function or Gaussian function, and interpolate peak position from the obtained function. Besides the function approximation, the center of gravity position may be obtained using the contrast value around the peak, and the obtained center of gravity position may be set as the peak position.

Here, if the peak position obtained by interpolation is A and the Z stage coordinate when the image of number 0 is taken is z0, the height h (x, y) of the peak position P is h (x, y). = Z0 + A × (λ / 8).

In the third embodiment, the same effect as in the first embodiment can be obtained.

[Embodiment 4]
The fourth embodiment relates to a method for increasing the detection accuracy of the height detection method of the third embodiment. First, problems of the third embodiment will be described.

If the wavelength of the light source is λ, the intensity value g _λ of the interference fringe waveform can be expressed by the following equation (7).

Here, s is the sampling position, h is the height of the ink application part, and α and γ are coefficients determined from the amplitude of white light. Since white light actually has a certain bandwidth, the center wavelength is λ, and light having a wavelength of λ1 ≦ λ ≦ λ2 is irradiated. The intensity of light having this bandwidth is expressed by the following equation (8).

Here, g _λ is given by Equation (7). G is one in which the wavelength lambda is varied between from .lambda.1 .lambda.2 adds g _lambda, and averaged by dividing by the number of times N obtained by adding.

In Expression (7), when s = h, that is, when the optical path length of the reflected light from the substrate 7 and the optical path length of the reference light from the reference mirror 32 are the same, cos (2π (2s−2h) / λ) is maximized and g _λ also takes a maximum value. Since this is the same at any wavelength, like the g _lambda G also takes the maximum value.

FIG. 13 is a diagram showing the relationship between the sampling position s and the interference fringe intensity G. FIG. The interference fringe intensity G in FIG. 13 is calculated using Equation (8). ● indicates sampling points. The sampling point controls the Z stage 8 that adjusts the relative position between the substrate 7 and the Mirau interference objective lens 30, and the relative distance between the substrate 7 and the objective lens 30 is λ / 8 corresponding to π / 2 in phase increment. (nm) is a plot of the luminance value G at the position (x, y) on the image when the image is taken while being changed. Note that the sampling of the image satisfies the Nyquist principle, and the original signal can be reproduced using the sampling points.

In Embodiment 3, the contrast value Mi of the interference fringe waveform is obtained from the luminance value of the sampling point, and the peak position is taken as the height of the corresponding pixel. The contrast value Mi is calculated by Equation (6) using a total of five images including two images before and after an image sample to be obtained when an image is taken while changing by π / 2 in phase increment.

½ of the square root of Mi corresponds to the envelope of the interference fringe waveform. When the envelope is superimposed on FIG. 13, the result is as shown in FIG. In actual measurement, α and γ in Equation (7) do not become constant due to the influence of noise, so the actual sampling point does not match the interference fringe waveform, and finally this mismatch is the position of the peak point. Deviation occurs.

[Advantages of using phase]
The phase information can be obtained without the influence of α and γ in Equation (7). Here, in order to make the explanation easy to understand, light having a center wavelength λ is considered. When the phase 2π (2s−2h) / λ of the interference fringe waveform is δ, Equation (7) becomes g _λ = α (1 + γ cos δ). Here, the following equation (9) is obtained from Euler's formula.

Here, when Expression (9) is Fourier transformed, and only the spectrum of the second term on the right side is extracted by a band pass filter and inverse Fourier transformed, the following Expression (10) is obtained.

The following equation (11) is obtained when the equation (10) is expressed by a trigonometric function according to Euler's formula.

Here, it can be seen that the phase δ is expressed by the following equation (12) and can be calculated without being affected by α and γ.

[Relationship between peak point and phase]
By the way, since the reflected light and the reference light pass through different optical paths before joining again, strictly speaking, it is preferable to consider the phase difference in Expression (7). This is because the reflection characteristics of the reference mirror 32 and the surface of the substrate 7 are different. When this phase difference is φ, the following equation (13) is obtained.

Here, the effect of the phase difference φ will be verified. FIG. 15 is a diagram showing the relationship between the sampling position s, the contrast value Mi, and the phase δ when the height h = 0 and the phase difference φ = 0. The horizontal axis in FIG. 15 indicates the sampling position s, the convex curve indicates Mi, and the saw-tooth line segment indicates the phase δ. The phase δ changes linearly downward from π to −π, and becomes discontinuous where it changes from −π to π. This discontinuous portion is indicated by a vertical line segment. Similarly to FIG. 13, an image is taken while changing the relative distance between the substrate 7 and the Mirau interference objective lens 30 by λ / 8 (nm) corresponding to π / 2 in phase increment.

As can be seen from FIG. 15, Mi reaches a peak at s = 0, and the phase becomes zero. That is, the peak point of the contrast value Mi coincides with the zero point of the phase δ at s = 0.

FIG. 16 is a diagram showing the relationship between the sampling position s, the contrast value Mi, and the phase δ when the height h = 0 and the phase difference φ = π / 2, and is compared with FIG. . The peak point of the contrast value Mi does not change and Mi is a peak point when s = 0, but the zero point of the phase δ has moved to the right as compared to FIG. The amount of movement is equal to the phase difference φ and is λ / 8. λ / 8 corresponds to π / 2 in phase increment. FIG. 16 shows that the peak of the contrast value Mi is not affected by the phase difference φ, but the zero point of the phase δ is affected by the phase difference φ.

[Combination of peak point and phase]
Although the peak of the contrast value Mi may be displaced due to noise, it can indicate the height of the object without being affected by the phase difference φ as described above. Theoretically, the phase δ can minimize the influence of noise, and can be detected with higher accuracy than the peak of the contrast value Mi. Therefore, in the present invention, the height of the ink application part is detected using both the peak of the contrast value Mi and the phase δ.

[Generation of phase jump]
Since it is difficult to obtain the phase difference φ in advance, the zero point of the phase δ closest to the peak of the contrast value Mi is set as the height of the corresponding pixel as an initial value. The peak of the contrast value Mi is called the primary height, and the zero point of the phase δ closest to the peak of the contrast value Mi is called the secondary height.

As can be seen from FIG. 15 and FIG. 16, there is one zero point of the phase δ on the left and right sides of the peak. In order to obtain the secondary height, the nearest zero point is adopted as the peak point. However, if the peak position is shifted due to noise, a zero point selection error is caused. When this selection error occurs, a phase jump of 2π occurs between adjacent pixels because the phase δ takes a value of −π to π.

FIG. 17 is a diagram showing a peak position of a certain line on the image. FIG. 17 shows data for one line in the horizontal direction when an inclined plane is measured. In FIG. 17, the horizontal axis indicates the pixel position, and the vertical axis indicates the image sampling number. It increases as the sampling number increases. When the sampling number changes by 1, the height changes by λ / 8.

Also, FIG. 18 shows the zero point of the phase δ closest to the peak position shown in FIG. Phase jumps occur at D and E in FIG. Also, comparing FIG. 17 with FIG. 18, it can be seen that there is less variation at the zero point of the phase δ.

[Detection and correction of phase jump]
Since the cause of the phase jump is a selection error of the zero point of the phase δ on the left and right of the peak point of the contrast value Mi, finally, either one of the left and right is selected uniformly in any pixel. It should be corrected as follows.

Therefore, as post-processing, the amount of deviation between the peak point of the contrast value Mi and the zero point of the phase δ is obtained, and correction processing is performed so that the sign of the amount of deviation coincides for almost all pixels on the image. To correct. Although this process changes the pixel height, there is no problem in this inspection method because the relative height is used for height evaluation.

First, the amount of deviation between the peak point and the zero point of the phase δ is Δ, the threshold is T, the threshold correction amount is t, and the number of corrections is M. If the peak of the contrast value Mi at the position (x, y) on the image is J (x, y) and the zero point of the phase is K (x, y), the deviation amount Δ (x, y) is Δ ( x, y) = K (x, y) −J (x, y).

Here, Δ (x, y) at the same place as in FIG. 17 is shown in FIG. The horizontal axis in FIG. 19 indicates the pixel position, the vertical axis indicates Δ (x, y), and a change of value 1 corresponds to λ / 8. Next, Δ (x, y) is compared with the threshold value T. If Δ (x, y) <T, K (x, y) is set to K ′ (x, y) = K (x, x y) Set to + 2π. If Δ (x, y)> T, K ′ (x, y) = K (x, y).

Thereafter, K ′ (x, y) after correction of all the pixels (x, y) is compared with at least one pixel adjacent to (x, y) to obtain a sum S of difference values. The difference value is the absolute value of the difference from K ′ (x, y) of the adjacent pixel. For example, the difference value from (x + 1, y) is | K ′ (x, y) −K ′ (x + 1, y) |. Further, the obtained total value S is stored in association with the threshold value T. Next, the correction amount t is added to the threshold value T to obtain a new threshold value T. Again, K ′ (x, y) is obtained for all pixels (x, y) on the image, and the total value S is obtained.

The above processing is repeated M times. Finally, the minimum value of the obtained total value S is obtained, and K ′ (x, y) is obtained again using the threshold value T when the minimum value is indicated, and the obtained K Let ′ (x, y) be the tertiary height of each pixel.

FIG. 21 shows the final K ′ (x, y) in FIG. Further, Δ (x, y) at this time is shown in FIG. In FIG. 21, the threshold value T is started from −4, and the correction amount t is set to 0.1, and the correction is performed 80 times in total. The threshold T varies up to 4. FIG. 21 shows that the phase jump is corrected.

[Switch detection method]
The amount of ink applied per time varies depending on the viscosity of the ink. High-viscosity ink has a large surface tension, so it becomes thicker than low-viscosity ink. In many cases, the properties of the ink are clarified in advance by a sample test.

In addition, the criteria for determining the amount of ink applied varies depending on the pattern to be applied. For thin films such as flat panel displays and semiconductors, the submicron level is required. It is.

Since the required detection accuracy varies depending on the ink to be applied and the pattern to be applied, the height information to be used can be switched according to the ink to be applied.

The ink to be applied can be changed for each application needle when, for example, the ink application mechanism 5 shown in FIG. 2 is used. Therefore, a selection column of “height type” is provided in the “needle-inspection item correspondence table” shown in FIG. 22, and the height type used when applying with the corresponding needle is registered. For example, when applied with the application needle A, the tertiary height is used, and when applied with the application needle B, the primary height is used.

Also, since the film thickness is different for each ink, the scan range can be set for each application needle. As a result, a tact time suitable for ink can be set, and the inspection time can be made more efficient.

In the fourth embodiment, the height of the ink application part can be detected with higher accuracy than in the third embodiment.

The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

1 defect correction device, 2 observation optical system, 2a observation barrel, 3 CCD camera, 4 cutting laser device, 5 ink application mechanism, 6 ink curing light source, 7 liquid crystal color filter substrate, 8 Z stage, 9 X stage, 10 Y stage, 11 control computer, 12 monitor, 13 operation panel, 15 movable plate, 16 objective lens, 17 coating unit, 18 coating needle, 19 ink tank, 20 electrostatic inkjet device, 21 inkjet nozzle, 22 pulse voltage generation Device, 23 control device, 24 correction ink, 24a tailor cone, 24b jet flow, 24c droplet, 30 Mirau interference objective lens, 31 lens, 32 reference mirror, 33 beam splitter, 34 falling light source, 35 filter switching device, 36 Motor, 37 a half mirror, 38 an imaging lens, 51 a black matrix, 52 R pixel, 53 G pixel, 54 B pixel, 55 white defect, 56 black defect, 57 defective foreign matter.

Claims

An application device for applying a liquid material to the surface of a substrate,
An observation optical system for observing the surface of the substrate via an objective lens, an imaging device for capturing an image of the surface of the substrate via the observation optical system, and an application mechanism for applying the liquid material to the surface of the substrate A head portion including
A positioning device that relatively moves the head portion and the substrate to position the head portion at a desired position above the surface of the substrate;
After controlling the positioning device and the imaging device and positioning the objective lens above the application portion made of the liquid material applied to the surface of the substrate, the application portion and the objective lens are relatively moved in the vertical direction. A coating apparatus comprising: a height detection unit that captures an image while moving, obtains a focal position for each of a plurality of pixels constituting the captured image, and obtains the height of the application unit based on the obtained focal position .
The imaging device captures an image of the application unit,
The positioning device includes a Z stage that mounts the head portion and moves in a direction perpendicular to the surface of the substrate,
The coating apparatus according to claim 1, wherein the height detection unit sets a position of the Z stage when a luminance contrast value of each pixel is maximized as a focal position of the pixel.
The objective lens is a two-beam interference objective lens;
The imaging device captures an image of interference fringes,
The positioning device includes a Z stage that mounts the head portion and moves in a direction perpendicular to the surface of the substrate,
The coating apparatus according to claim 1, wherein the height detection unit sets the position of the Z stage when the contrast value of the interference fringe intensity of each pixel is maximized as a focal position of the pixel.
The coating apparatus according to claim 3, wherein the height detection unit samples a captured image at a predetermined cycle and obtains a focal position of each pixel of the sampled image.
The coating apparatus according to claim 4, wherein the predetermined period is a period that satisfies a Nyquist period.
The height detector calculates the contrast value of the interference fringe intensity for each pixel of the sampled image, sets the sampling position of the image where the calculated value reaches a peak as the primary height, and further obtains the phase from the interference fringe intensity. Then, the nearest zero point to the primary height is set as the secondary height of the corresponding pixel, and the occurrence of phase jump is detected for each pixel by comparing the primary height and the secondary height. The coating apparatus according to claim 4, wherein the secondary height is corrected when the detection is detected, and the tertiary height with the phase jump corrected is used as a measurement result.
The height detector is
When determining the tertiary height, a threshold value is provided, and for all pixels on the image, the value obtained by subtracting the primary height from the secondary height of each pixel is compared with the threshold value. When the subtraction result is less than the threshold value, the sampling position whose phase has increased by 2π with respect to the current secondary height is taken as the corrected secondary height, and each pixel for the corrected secondary height Every time, the difference value between adjacent pixels is calculated to obtain the sum of the difference values of all the pixels on the image,
In this process, the sum of the difference values is held for each threshold value while correcting the threshold value a total of M times in a predetermined correction amount, and the sum of the corrected secondary heights is minimized. The coating apparatus according to claim 6, wherein a secondary height corrected by a threshold value is set as the tertiary height.
The coating apparatus according to claim 6, wherein the height detection unit selects and measures either the primary height or the tertiary height according to the coating condition of the liquid material.
The coating apparatus according to claim 1, further comprising an inspection unit that determines whether the coating unit is normal based on a height of the coating unit detected by the height detection unit.
The inspection unit subtracts the height of the reference portion on the surface of the substrate on which the liquid material is not applied from the height of the application unit detected by the height detection unit. The coating apparatus according to claim 9, wherein a height is obtained, and whether or not the coating unit is normal is determined based on the relative height.
The inspection unit is normal based on at least one of a total value, a maximum value, a minimum value, an average value, and a dispersion value of the relative heights of the plurality of pixels constituting the image of the application unit. The coating apparatus of Claim 10 which determines whether it is.
The coating apparatus according to claim 1, wherein the coating mechanism applies the liquid material adhering to a tip portion of a coating needle to the surface of the substrate.
The coating apparatus according to claim 1, wherein the coating mechanism applies the liquid material to a surface of the substrate by an electrostatic ink jet method.
An observation optical system for observing the surface of the substrate via an objective lens, an imaging device for capturing an image of the surface of the substrate via the observation optical system, and applying the liquid material to the surface of the substrate A coating apparatus comprising: a head unit including a coating mechanism; and a positioning device that relatively moves the head unit and the substrate to position the head unit at a desired position above the surface of the substrate. A height detection method for detecting the height of the application part made of the liquid material applied to the surface of
After the positioning device and the imaging device are controlled and the objective lens is positioned above the application unit, an image is captured while the application unit and the objective lens are relatively moved in the vertical direction. A height detection method for obtaining a focal position for each of a plurality of constituent pixels and obtaining a height of the application unit based on the obtained focal position.