JP4650096B2 - Coating unevenness inspection method and program thereof - Google Patents

Description

  The present invention relates to a method for detecting uneven application of a transparent resin film applied in a certain direction on a substrate. More specifically, the present invention relates to a detection method suitable for detecting uneven application of a transparent photosensitive resin film for a photospacer used for manufacturing a liquid crystal display and a program thereof.

  A liquid crystal display is configured by adhering two electrode plates each having an electrode to each other with a gap therebetween and enclosing liquid crystal in the gap. A voltage is applied between the electrodes for each pixel to drive the liquid crystal and control the polarization plane of the light passing through the liquid crystal to control the transmission / non-transmission of the polarizing film and display the screen. .

  Conventionally, the gap is usually maintained by the diameter of the spherical beads dispersed on one electrode plate. However, the dispersion position of the spherical beads cannot be controlled. For this reason, display unevenness may occur depending on the dispersion position.

  In order to solve this, in recent years, the gap is controlled by a large number of photo spacers. The photospacer is a columnar spacer provided on one of the electrode plates, and is generally provided by exposing and developing a transparent photosensitive resin applied to the electrode plate. In the case of a color liquid crystal display device, a black frame that divides a display screen on a glass substrate, a black light shielding film (black stripe or black matrix) that divides the inside of the frame into pixels, and the light shielding film. In addition, a colored film that colors display light is provided in each pixel, and a transparent photosensitive resin film is applied to cover the black frame, the light shielding film, and the colored film. According to this method, since the position of the photospacer can be specified by the exposure mask, there is no fear of display unevenness caused by the conventional spherical beads.

  However, since the height of these columnar photospacers is determined by the thickness of the applied photosensitive resin film, if the photosensitive resin film has uneven application, the heights of these photospacers are also uneven. The distance between the electrodes becomes unstable and the thickness of the liquid crystal layer becomes unstable, which affects the display quality. For this reason, it is necessary to apply the photosensitive resin film to a uniform film thickness, and on the other hand, it is necessary to detect a photosensitive resin film having uneven coating before exposure and development, and to peel and remove this film. If a photo spacer is formed by exposing and developing a photosensitive resin film with uneven coating, it is difficult to detect the unevenness of the height of the photo spacer, and it is also difficult to peel and remove from the electrode plate. That's why.

  In recent years, liquid crystal displays have a tendency to lower the price due to market competition, and in order to cope with this, electrode plates are manufactured by attaching a large number of glass plates to a large area. In addition, since the display screen of the liquid crystal display itself is becoming larger, the glass substrate used for this is becoming more and more likely to have a larger area. It is becoming.

Usually, since the transparent photosensitive resin used for the photo spacer of a liquid crystal display needs to apply | coat uniformly on a large area board | substrate, it is apply | coated to the fixed direction by the die coater. In this way, the transparent photosensitive resin applied in a certain direction often has application unevenness in a straight line. For example, when the slit width for extruding the transparent photosensitive resin is partially changed during coating with this die coater, coating unevenness occurs linearly along the coating direction. Further, when the feed accuracy of the die coater fluctuates, linear coating unevenness occurs in a direction orthogonal to the coating direction.
In addition, even if there is a foreign object between the die coater stage and the glass substrate, the film thickness varies, causing irregularities in round and irregular shapes, or due to the temperature difference between the support pins supporting the glass substrate and the glass substrate. Dry pin unevenness may occur.

  In general, when light is applied to a transparent coating, reflected light having an intensity different from that of the surrounding area (normal part) is observed at a part where uneven application occurs (uneven part). For example, when observation is performed using interference light, the intensity of reflected light is generated by a change in film thickness due to the relationship between the wavelength λ of the light and the film thickness. If the phase of the film thickness and the wavelength λ match, they will strengthen each other, and if the film thickness changes and the phase shifts, the interference light becomes weak. For this reason, the coating unevenness can be observed with the naked eye.

  It is also possible to detect coating unevenness by observing and measuring the intensity of the reflected light with a camera or the like. However, the sensitivity of a camera or the like is lower than that of the naked eye, and it is difficult to distinguish between a normal part and an uneven part. In addition, there is almost no problem with visual inspection, and it is caused by minute film thickness fluctuations that fall within the allowable range. The influence of a lattice or the like existing in the lower layer of the photosensitive resin film becomes obvious in the image. If there is this influence, the uneven portion cannot be easily detected by simple image processing means such as binarization processing or filtering processing.

  Attempts have been made to improve this, increase the S / N between the normal portion and the uneven portion, and detect the coating unevenness with high accuracy. For example, the region to be inspected is divided into pixels, irradiated with two types of light with different wavelengths, the reflected light intensity is measured for each pixel and for each wavelength, and then the difference in reflected light intensity based on the wavelength is calculated. Calculate and accumulate the difference for a predetermined length in the first direction, differentiate this accumulated value in the first direction, and subtract the minimum value from the maximum value of the differential value between the predetermined sections in the second direction Although there is a method of calculating the difference and comparing the differential difference with a threshold value to detect coating unevenness (see Patent Document 1), it is difficult to perform unevenness inspection that takes into account the effects of the above-mentioned moire unevenness and the lattice existing in the lower layer. Met. Further, the processing speed is not satisfactory in terms of detecting and processing two types of optical signals.

Japanese Patent Laid-Open No. 2003-166941

  The present invention has been made under the circumstances as described above. The reflected light intensity is accurately measured on the entire substrate having a large area, and the lattice existing under the transparent photosensitive resin film is determined from the reflected light intensity. An object of the present invention is to provide a method for efficiently detecting a coating unevenness of a transparent photosensitive resin coating by accurately estimating a region (unevenness candidate region) where there is a possibility of coating unevenness while removing disturbance factors such as influences. To do.

  That is, the invention according to claim 1 is a method for inspecting application unevenness of a transparent resin film applied on a substrate, and is divided into a plurality of small regions by dividing the application direction of the transparent resin and a direction perpendicular thereto. The reflected light intensity measuring step for irradiating light to these small areas and measuring the reflected light intensity for each small area, and the coating direction or the direction orthogonal to the coating direction as the envelope direction, are arranged along this envelope direction. An envelope value calculating step of calculating the envelope value by enveloping the reflected light intensity of the small area, and the envelope value arranged along this integration direction, with the application direction or the direction orthogonal to the application direction as the integration direction. A non-uniformity inspecting method comprising: a non-uniformity extracting step of integrating and extracting a region where the integrated envelope value is larger or smaller than a surrounding integrated value as a region where there is a possibility of non-uniformity in coating It is.

  According to this method, the transparent photosensitive resin is obtained by calculating the envelope value by performing the envelope processing from the reflected light intensity of the small region arranged along the envelope direction, with the coating direction or the direction orthogonal to the coating direction as the envelope direction. Since the lattice effect existing in the lower layer of the coating is removed, the application direction or the direction orthogonal to the application direction is set as the integration direction, and envelope values arranged along this integration direction are integrated to calculate the integrated value, When there is linear coating unevenness along the integration direction, the difference between the integrated value of the linear unevenness part and the integrated value of the normal part is larger than the difference between the reflected light intensities in a single small area. The detection accuracy can be improved.

  Next, the invention according to claim 2 is a method for inspecting the application unevenness of the transparent resin film applied on the substrate, and is divided into a direction in which the transparent resin is applied and a direction perpendicular thereto, and a number of small portions are divided. A reflected light intensity measuring step for irradiating light to these small areas and measuring the reflected light intensity for each small area, and the direction of application or the direction perpendicular to the application direction as the moving average direction. A moving average value calculating step for calculating a moving average value by moving average of reflected light intensities of small areas arranged along the line, and a coating direction or a direction perpendicular to the coating direction as an envelope direction, lined along the envelope direction An envelope value calculating step for calculating the envelope value by enveloping the reflected light intensity of the small area, and a non-uniformity extracting step for extracting coating unevenness by a comparison operation between the moving average value and the envelope value. How to check coating unevenness It is.

  According to this method, the moving average value and the envelope value are calculated from the reflected light intensity of the small regions arranged along the moving average direction and the envelope direction, with the application direction or the direction orthogonal to the application direction as the moving average direction and the envelope direction. Then, the lattice effect existing under the transparent photosensitive resin film is removed by obtaining the envelope value, and the brightness fluctuation (moy uneven effect) existing over the entire surface of the substrate is reduced by comparison with the moving average value. And the detection accuracy can be improved.

  Next, the invention according to claim 3 is a coating unevenness inspection method capable of more accurately removing the lattice effect existing in the lower layer of the transparent photosensitive resin film by obtaining the upper envelope value and the lower envelope value. is there.

  Next, the invention according to claim 4 performs brightness fluctuation (over the entire surface of the substrate) by dividing the upper envelope value and the moving average value and dividing the lower envelope value and the moving average value. The aim is to reduce the effect of mist. Further, the present invention is a coating unevenness inspection method in which a threshold value is provided in advance for the division value, and a portion showing a division value larger or smaller than the threshold value is extracted as a region where there is a possibility of coating unevenness.

  Next, the invention according to claim 5 divides the upper envelope value and the moving average value, and divides the lower envelope value and the moving average value, and further multiplies each divided value, that is, the upper envelope. Aimed at emphasizing coating unevenness by synergizing brightness fluctuations occurring in the upper envelope value indicating the bright part and the lower envelope value indicating the dark part by calculating the ratio of the value and the lower envelope value It is. Further, the present invention is a coating unevenness inspection method in which a threshold value is previously set for the multiplication value, and a portion showing a multiplication value larger or smaller than the threshold value is extracted as a region where there is a possibility of coating unevenness.

  Next, the invention according to claim 6 divides the transparent resin applied on the base material into a coating direction and a direction perpendicular thereto to form a large number of small regions, and the reflected light of the light irradiated to these small regions This is a program for inspecting coating unevenness from the strength, and the envelope value is obtained by enveloping the reflected light intensity of a small area aligned along the envelope direction with the coating direction or the direction orthogonal to the coating direction as the envelope direction. And the application direction or a direction perpendicular to the application direction as an integration direction, the envelope values arranged along the integration direction are integrated, and the integrated envelope value is compared with the surrounding integrated value. And a step of extracting a large or small portion as a region having a possibility of uneven coating.

  According to this method, the transparent photosensitive resin is obtained by calculating the envelope value by performing the envelope processing from the reflected light intensity of the small region arranged along the envelope direction, with the coating direction or the direction orthogonal to the coating direction as the envelope direction. Since the lattice effect existing in the lower layer of the coating is removed, the application direction or the direction orthogonal to the application direction is set as the integration direction, and envelope values arranged along this integration direction are integrated to calculate the integrated value, When there is linear coating unevenness along the integration direction, the difference between the integrated value of the linear unevenness part and the integrated value of the normal part is larger than the difference between the reflected light intensities in a single small area. Thus, it is possible to perform processing for improving detection accuracy.

  Next, the invention according to claim 7 divides the transparent resin applied on the base material into a coating direction and a direction perpendicular thereto to form a large number of small regions, and the reflected light of the light irradiated to these small regions This is a program for inspecting coating unevenness based on the intensity. The moving average direction is defined as the moving average direction in the coating direction or the direction orthogonal to the coating direction. A step of calculating a moving average value, a step of calculating an envelope value by enveloping the reflected light intensity of a small region arranged along the envelope direction with the application direction or a direction orthogonal to the application direction as an envelope direction; And a step of extracting coating unevenness by comparison between the moving average value and the envelope value.

  According to this method, the moving average value and the envelope value are calculated from the reflected light intensity of the small regions arranged along the moving average direction and the envelope direction, with the application direction or the direction orthogonal to the application direction as the moving average direction and the envelope direction. Then, the lattice effect existing under the transparent photosensitive resin film is removed by obtaining the envelope value, and the brightness fluctuation (moy uneven effect) existing over the entire surface of the substrate is reduced by comparison with the moving average value. Therefore, it is possible to perform processing for improving detection accuracy.

  In general, data with a lot of noise changes abruptly, so that the effect of the noise can be suppressed by further smoothing the data. Further, the difference between the data can be amplified by adding the surrounding data to the specific data and replacing the specific data with the added value.

  According to the present invention, the reflected light intensity is accurately measured on the entire substrate having a large area, and the brightness existing over the front surface of the substrate due to the minute film thickness fluctuation belonging to the allowable range is calculated from the reflected light intensity. It is possible to remove disturbance factors such as the influence of fluctuation (moyamura) and the lattice effect existing under the transparent photosensitive resin film, and accurately estimate the area (unevenness candidate area) that may have coating unevenness. As a result, it is possible to accurately detect coating unevenness.

  The present invention uses the present invention, for example, with an intermediate product applied to the production of an electrode plate for a liquid crystal display as a base material, and a transparent photosensitive resin film coated on the base material with a die coater as a transparent resin film. be able to.

  FIG. 1 is an explanatory plan view showing the intermediate product 1 and the transparent photosensitive resin applied to the surface of the intermediate product 1, and illustrations of points unnecessary for the description are omitted.

  That is, the intermediate product 1 is partitioned on a glass substrate by a black frame 11 that partitions the display screen 12, a black matrix (not shown) that partitions the frame 11 into pixels, and the black matrix. In addition, a color film (not shown) that is provided in each pixel and colors display light is provided. In addition, although it is normal that the display screen is multi-faced on one glass substrate, only one display screen is shown in the figure.

  A transparent photosensitive resin is applied on the intermediate product with a die coater. The transparent photosensitive resin is applied to a uniform film thickness by starting application from the glass substrate end face 13 and conveying the slit die of the die coater or conveying the glass substrate in the illustrated direction. In the figure, reference numeral 14 denotes an application region, and 15 denotes application unevenness. The coating unevenness shown in the figure is due to pin unevenness due to foreign matter and unevenness of the gap between the slit die of the die coater and the substrate. The former is round and indefinite in any place, the latter is the coating direction or the coating orthogonal direction Along a straight line.

  And this invention can be implemented using an apparatus as shown, for example in FIG. When performing using the apparatus of FIG. 2, first, the glass substrate 1 coated with the transparent photosensitive resin is placed and fixed on the stage including the gantry 4.

  Any stage may be used, but it is desirable to use a stage with a black surface in order to prevent reflected noise light from the stage surface. For example, an aluminum stage whose surface is anodized to be black. Or it is good also considering only the glass substrate 1 lower part in an imaging point part as a surface black member.

  The glass substrate 1 can be placed and fixed by, for example, providing a vacuum suction mechanism on the upper surface of the stage and sucking it by this mechanism. Moreover, you may fix to the hollow position above a stage using a lifter, and it is also possible to fix to the hollow position above a stage by gripping the end surface of the glass substrate 1 with a clamp. By fixing in the hollow position in this way, reflected noise light from the stage surface can be prevented.

  2, the gantry 4 is equipped with a plurality of light sources 2 and a camera 3 that receives reflected light obtained by reflecting light emitted from the light sources with a transparent resin film.

  As the camera 3, a monochrome camera can be preferably used, but a color camera may be used. The camera 4 may be either a line camera or an area camera. Note that it is preferable to use a high-sensitivity camera of 10 bits or more.

  The camera 3 is preferably arranged so that its optical axis forms an angle of 40 to 60 ° with respect to the normal line of the substrate 1 in terms of measurement sensitivity. More desirably, it is 50 °. Further, it is desirable that the light source 2 is also arranged so that the irradiation light is incident at an angle of 40 to 60 ° with respect to the normal line of the substrate 1 corresponding to the angle of the camera 3.

  Then, the light from the light source 2 is irradiated, and the intensity of the light reflected by the transparent photosensitive film on the substrate 2 is measured by the camera 2. By measuring the reflected light intensity by moving the gantry 4, it is possible to measure the reflected light intensity for the entire measurement target region. If the measurement range is wide, the substrate 2 itself is moved and the gantry 4 is moved if there is not enough room for installation, such as when the intermittent movement of the gantry and the measurement of the reflected light intensity are repeated or provided on the in-line. The reflected light intensity measurement of the entire range may be performed in a state in which is stopped.

  The reflected light intensity is divided into a plurality of small regions (m, n) by dividing the transparent resin in the direction of application (m direction) and the direction (n direction) perpendicular thereto. It is necessary to irradiate light and measure the reflected light intensity R for each small area (m, n). In the apparatus of FIG. 2, the moving direction of the gantry 4 or the moving direction of the glass substrate is defined as the coating direction (m direction), but the moving direction of the gantry 4 or the moving direction of the glass substrate is orthogonal to the coating direction (m direction). You may arrange | position so that it may become a direction (n direction).

  The small area (m, n) may be an area corresponding to each pixel of the camera 3. For this reason, the higher the resolution of the camera 3, the more the surface of the transparent photosensitive resin coating is divided into fine small regions, which enables accurate measurement. It is also possible to combine a plurality of pixels of the camera 3 to correspond to the small area (m, n).

  Next, the light used for the inspection of film thickness unevenness in the present invention may be light of any wavelength, but when the transparent photosensitive resin film has photosensitivity, it is not sensitive to the transparent photosensitive resin film. It is desirable to use light of a wavelength. In general, since a photosensitive resin is sensitive to short-wavelength light such as ultraviolet rays, in the present invention, green light belonging to a wavelength range of 500 to 570 nm, or 580 to 630 nm is used as light for inspection of film thickness unevenness. Red light belonging to the wavelength region can be suitably used. These lights are preferably narrow-band rays. Preferably, it is monochromatic light with a half-width of 20 nm or less.

  Green monochromatic light belonging to a wavelength range of 500 to 570 nm can be obtained by using, for example, a halogen lamp or a high-pressure mercury lamp as the light source 2 and transmitting the light through a band-pass filter. Red monochromatic light can be obtained by using a halogen lamp or a low-pressure sodium lamp as the light source 2 and transmitting the light through a band-pass filter. In addition, it is necessary to pay sufficient attention to measures against heat to the bandpass filter and atmosphere.

  In addition, the irradiation light with which the transparent photosensitive resin is irradiated may include light belonging to another wavelength region in addition to the light used for inspection. In this case, the reflected light intensity of the light used for the inspection can be measured by passing the reflected light through the band pass filter. Further, after measuring the reflected light intensity for each color light with a color camera, the reflected light intensity of the light used for the inspection may be extracted on the data.

  Next, the moving average value and the envelope value are calculated from the reflected light intensity R (m, n) of the small regions (m, n) arranged along the coating direction (m direction) or the direction orthogonal to the coating direction (n direction). calculate. The moving average value and the envelope value are calculated as means for removing the lattice effect existing in the lower layer of the transparent photosensitive resin, and the direction orthogonal to the pixel stripe formed on the glass substrate is preferable. This is because in the direction perpendicular to the pixel stripes, the lattice effect existing in the lower layer of the transparent photosensitive resin is noticeably generated in the lattice arrangement structure, and the undulation (moire) generated over the entire substrate appears more accurately. This is because it is suitable for generation. The direction of the pixel stripe varies depending on the imposition arrangement on the glass substrate, and there are two types of cases: the same direction as the application direction (m direction) and the same direction as the direction orthogonal to the application direction (n direction). Yes, choose timely.

  Hereinafter, a case where the moving average value and the envelope value are calculated from the reflected light intensity R (m, n) of the small regions (m, n) arranged in the application direction (m direction) will be described as an example (application direction (m direction). ) In a direction perpendicular to the pixel stripe).

  The moving average value obtained from the reflected light intensity R (m, n) of the small regions (m, n) arranged in the coating direction (m direction) is, for example, the reflected light intensity Rm (m, n) and Rm + 1 (adjacent to each other). m, n) and calculating the average. For example, (R1 + R2) / 2, (R2 + R3) / 2,. Of course, not only the two reflected light intensities Rm adjacent to each other but also an average of three or more reflected light intensities Rm that are continuous with each other can be used as the moving average value Rm ′. As the smoothing process, it is also possible to use a process of removing a high-frequency component after Fourier transform and performing inverse transform.

  The envelope value obtained from the reflected light intensity R (m, n) of the small regions (m, n) arranged in the coating direction (m direction) can be obtained by the following method, for example. The difference between the reflected light intensity Rm and the moving average value Rm ′ is obtained. At this time, a positive value is an upper envelope value candidate, and a negative value is a lower envelope value candidate, but these values are discrete values. Therefore, the discrete value is replaced with a continuous value. In the case of the upper envelope value, the maximum value of the envelope value candidate is obtained in a specific interval (m, n) − (m + m ′, n), and (m + m ′, n) − (m + 2m ′, n). In the same way, the maximum value of the envelope value candidates is obtained. By interpolating between the maximum values of the envelope value candidates obtained in each section by linear approximation interpolation, the gap data is compensated, and data for tracing the bright part side of the reflected light intensity, that is, the upper envelope value Tm is obtained. In the case of the lower envelope value, the minimum value of the envelope value candidate is obtained in a specific section (m, n) − (m + m ′, n), and (m + m ′, n) − (m + 2m ′, n). Similarly, the minimum value of the envelope value candidates is obtained in the interval of. By interpolating between the minimum values of the envelope value candidates obtained in each section by linear approximation interpolation, the gap data is compensated, and data for tracing the dark side of the reflected light intensity, that is, the lower envelope value Bm is obtained.

  As another envelope data generation method, the maximum value and the minimum value of the reflected light intensity R are obtained in a certain section, and the maximum value and the minimum value in each section are similarly obtained while sequentially moving the section. Next, it is possible to generate the upper envelope value Tm by interpolating between the maximum values of the respective sections obtained, and similarly interpolating between the minimum values of the sections. For the section, the period can be predicted and set from the arrangement structure such as the pitch of the underlying pixel section.

  By performing such an envelope process, the reflected light intensity is re-created as gentle data (image) in the same way as observed by the human eye, aiming to improve the detection sensitivity of the required coating unevenness. The aim is to improve the detection accuracy of coating unevenness by reducing the influence of disturbance by reducing high-frequency fluctuations present in the reflected light intensity data, that is, the lattice effect existing in the lower layer of the transparent photosensitive resin.

  Further, brightness fluctuations (Hyamura) existing on the entire surface of the substrate by division Tm / Rm ′ between upper envelope value Tm and moving average value Rm ′ and division Bm / Rm ′ between lower envelope value Bm and moving average value Rm ′. Aim to reduce the impact). The division value between the upper envelope value and the moving average value is always 1 or more. On the contrary, since the division between the lower envelope value and the moving average value is always a value of 1 or less, for example, when threshold processing is to be performed with the same scale, the division value Rm of the moving average value and the lower envelope value is used. It may be '/ Bm. This makes it possible to quantitatively determine whether or not there is coating unevenness, which is a sensory defect.

  Further, a threshold value Tht is provided in advance for the division value Tm / Rm ′ between the upper envelope value Tm and the moving average value Rm ′, and a threshold value Thb is provided for the division value Bm / Rm ′ between the lower envelope value Bm and the moving average value Rm ′. It is possible to extract a region that is likely to be a coating unevenness in a portion that shows a division value larger or smaller than the threshold values Tht and Thb.

  Further, the obtained division values Tm / Rm ′ and Bm / Rm ′ or Rm ′ / Bm are reconstructed as images, and image processing such as filtering processing is performed on these images, which may cause coating unevenness. Region extraction may be attempted. In this image, since the image is reconstructed by the division value obtained from the reflected light intensity R as a base, the brightness fluctuation and the lower layer lattice effect existing on the entire surface of the substrate are reduced, and the image processing is simpler than the base data. Thus, uneven coating can be extracted, and the inspection speed can be increased as a whole.

Also, the product Tm / Rm ′ × multiplied by multiplying the division value Tm / Rm ′ of the upper envelope value Tm and the moving average value Rm ′ by the division value Rm ′ / Bm of the moving average value and the lower envelope value. By obtaining Rm ′ / Bm = Tm / Bm and synergizing brightness variations occurring in the upper envelope value indicating the bright portion and the lower envelope value indicating the dark portion, it is possible to enhance the coating unevenness.
Further, it is possible to provide a threshold value Thtb for the multiplication value Tm / Bm, and to extract a region having a possibility of application unevenness as a part showing a multiplication value larger than the threshold value Thtb.
Alternatively, the obtained multiplication value Tm / Bm may be reconstructed as an image, and image processing such as filtering processing may be performed on these images to try to extract a region that is likely to be coating unevenness. In this image, since the image is reconstructed by the division value obtained from the reflected light intensity R as a base, the brightness fluctuation and the lower layer lattice effect existing on the entire surface of the substrate are reduced, and the upper envelope value and the lower envelope value By using the multiplication value, the coating unevenness is more emphasized, so that the coating unevenness can be extracted by simple image processing from the base data, and the inspection accuracy can be improved.

  In this way, it is possible to detect circular and irregular coating unevenness due to foreign matters.

  In addition, the coating unevenness generated in a straight line is the upper envelope value Tm and the lower envelope value Bm obtained from the reflected light intensity R (m, n) of the small regions (m, n) arranged in the coating direction (m direction). Are integrated to calculate the integrated values ΣSmt and ΣSmb, the coating unevenness generated linearly along this direction (m direction) can be detected. For example, coating unevenness caused by partial fluctuation of the slit width of the die coater.

  Further, the upper envelope value Tn and the lower envelope value Bn obtained from the reflected light intensity R (m, n) of the small regions (m, n) arranged in the direction (n direction) orthogonal to the coating direction are integrated and integrated. When the values ΣSnt and ΣSnb are calculated, it is possible to detect coating unevenness that occurs linearly along this direction (n-direction). For example, coating unevenness caused by fluctuations in the feed accuracy of the die coater.

  Further, when the integrated value ΣSmt, ΣSmb, ΣSnt, ΣSnb is calculated by integrating the upper envelope value Tmn and the lower envelope value Bmn in each of the application direction (m direction) and the direction orthogonal to the application direction (n direction). Both coating irregularities can be detected.

Further, these integrated values ΣSm and ΣSn are compared with each other, and a portion that is larger or smaller than the surrounding integrated values ΣSm and ΣSn is extracted as a region where there is a possibility of uneven application, and the integrated values ΣSm and ΣSn have a threshold Ths. Can be extracted as a region where there is a possibility of coating unevenness, which indicates an integrated value ΣSm, ΣSn that is larger or smaller than the threshold Ths.
Such calculation can be performed by a computer (not shown).

  The outline of the substrate 1 used in this example is as shown in FIG. A transparent photosensitive resin sensitive to blue light is applied in one direction using a die coater, and there are two types of round coating unevenness: unevenness that looks white and unevenness that looks black. In this embodiment, these round unevenness will be described.

  Next, the outline of the apparatus used in this embodiment is as shown in FIG. As a light source, a low-pressure sodium lamp was provided in an illumination box, and the illumination box was disposed and diffused so that the central axis of the illumination box was 50 ° with respect to the normal of the substrate surface. As the camera 3, a black and white line camera was used, which was arranged so that its optical axis was 50 ° with respect to the normal of the substrate surface. And the board | substrate 1 was moved, the reflected light intensity | strength R was measured about the board | substrate whole area | region, and the moving average value R 'and the envelope values T and B were calculated | required from the reflected light intensity R.

  First, in the unevenness that appears white, an order m is given along the coating direction (m direction), the order m is a variable, and the moving average value Rm ′ is a function. Displayed. Further, with the upper envelope value Tm and the lower envelope value Bm as functions, these variables m and the functions Tm and Bm are displayed on orthogonal coordinates. These are shown in FIG.

  From FIG. 3, first, in the uneven region, the reflected light intensity Rm is finely fluctuated due to the influence of the lower layer grating compared to the surrounding area, and there is a large undulation (non-periodic fluctuation) from the moving average value Rm ′. It can be seen that the projection protrudes upward. At the time of visual observation, fine periodic fluctuations are not visible, and coating unevenness is pointed out from large undulations. Therefore, by calculating the gentle curve by eliminating the irregularities with the upper envelope value Tm and the lower envelope value Bm. It can be seen that the difference can be clearly recognized.

  Next, a division Tm / Rm ′ between the upper envelope value Tm and the moving average value Rm ′, a division Bm / Rm ′ between the lower envelope value Bm and the moving average value Rm ′, and a multiplication value Tm / Rm ′ × Rm ′. / Bm = Tm / Bm was determined and is shown in FIG.

  From FIG. 4, large undulation is reduced from the division values Tm / Rm ′ and Bm / Rm ′, and the flattening makes it easy to understand the presence of coating unevenness and simplifies the processing.

  Further, the obtained division values Tm / Rm ′ and Bm / Rm ′ or Rm ′ / Bm are reconstructed as images, and image processing such as filtering processing is performed on these images, which may cause coating unevenness. Region extraction may be attempted. In this image, since the image is reconstructed by the division value obtained from the reflected light intensity R as a base, the brightness fluctuation and the lower layer lattice effect existing on the entire surface of the substrate are reduced, and the image processing is simpler than the base data. Thus, uneven coating can be extracted, and the inspection speed can be increased as a whole.

  By the way, since this unevenness is unevenness that appears white, when the upper envelope value is compared with the lower envelope value, the existence of the upper envelope value is easier to understand. Therefore, by further calculating the ratio of the upper envelope value and the lower envelope value, which is finally the multiplication value Tm / Rm ′ × Rm ′ / Bm = Tm / Bm, the coating unevenness is further emphasized and the difference is clear. Can be recognized. Further, by providing a threshold value for this multiplication value, it is possible to discriminate between a site where unevenness exists and a normal site.

  As an example, for unevenness that appears black, an order m is given along the coating direction (m direction), the order m is a variable, and the moving average value Rm ′ is a function, and the variable m and the function Rm ′ are orthogonal to each other. Displayed on coordinates. Further, the variable m and the functions Tm and Bm are displayed on the orthogonal coordinates with the upper envelope value Tm and the lower envelope value Bm as functions. These are shown in FIG.

  From FIG. 5, first, in the non-uniform region, the reflected light intensity Rm is finely fluctuated due to the influence of the lower layer grating and there is a large undulation (non-periodic fluctuation) from the moving average value Rm ′. It can be seen that the protrusion protrudes downward. At the time of visual observation, fine periodic fluctuations are not visible, and coating unevenness is pointed out from large undulations. Therefore, by calculating the gentle curve by eliminating the irregularities with the upper envelope value Tm and the lower envelope value Bm. It can be seen that the difference can be clearly recognized.

  Next, a division Tm / Rm ′ between the upper envelope value Tm and the moving average value Rm ′, a division Bm / Rm ′ between the lower envelope value Bm and the moving average value Rm ′, and a multiplication value Tm / Rm ′ × Rm ′. / Bm = Tm / Bm was determined and is shown in FIG.

  From FIG. 6, large undulation is reduced from the division values Tm / Rm ′ and Bm / Rm ′, and the flattening makes it easy to understand the presence of coating unevenness and simplifies the processing. Alternatively, the obtained multiplication value Tm / Bm may be reconstructed as an image, and image processing such as filtering processing may be performed on these images to try to extract a region that is likely to be coating unevenness. In this image, since the image is reconstructed by the division value obtained from the reflected light intensity R as a base, the brightness fluctuation and the lower layer lattice effect existing on the entire surface of the substrate are reduced, and the upper envelope value and the lower envelope value are reduced. By using the multiplication value, the coating unevenness is more emphasized, so that the coating unevenness can be extracted by simple image processing from the base data, and the inspection accuracy can be improved.

  By the way, since this unevenness is unevenness that appears black, the presence of the lower envelope value is easier to understand when the upper envelope value and the lower envelope value are compared. Therefore, by further calculating the ratio of the upper envelope value and the lower envelope value, which is finally the multiplication value Tm / Rm ′ × Rm ′ / Bm = Tm / Bm, the coating unevenness is further emphasized and the difference is clear. Can be recognized.

  Further, by providing a threshold value for this multiplication value, it is possible to discriminate between a site where unevenness exists and a normal site. For example, when the threshold value is 1.3, it is possible to discriminate between a region where unevenness exists and a normal region with the same threshold value for both the white unevenness and the black unevenness.

  From these results, according to the present invention, from the comparative calculation value obtained from the reflected light intensity on the entire surface of the substrate, the lattice effect and the large waviness (moisture unevenness) in the lower layer of the transparent photosensitive resin are reduced, and the possibility of uneven application can be accurately detected. It was found that a certain area can be extracted and the inspection can be made more efficient.

  It is suitable for detecting coating unevenness of a transparent photosensitive resin film for a photospacer used for manufacturing a liquid crystal display.

It is a top view for description of the board | substrate which has transparent photosensitive resin. It is a perspective view for explanation of a detection device concerning the present invention. It is a graph which shows the reflected light intensity which concerns on the white nonuniformity example of this invention, a moving average value, and an envelope value. It is a graph which shows the comparison calculation value which concerns on the white nonuniformity Example of this invention. It is a graph which shows the reflected light intensity which concerns on the black nonuniformity example of this invention, a moving average value, and an envelope value. It is a graph which shows the comparison calculation value which concerns on the black nonuniformity Example of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Base material which provides a transparent photosensitive film 2 ... Light source 3 ... Camera 11 ... Black picture frame 12 ... Display screen 13 ... End surface 14 ... Application | coating area | region 15 ... Application | coating village

Claims (7)

In the method of inspecting the coating unevenness of the transparent resin film applied on the substrate, wherein the substrate has a lattice present in the lower layer of the transparent resin film,
A reflected light intensity measuring step of dividing the application direction of the transparent resin and a plurality of small areas by dividing the direction into the direction orthogonal thereto, irradiating light to these small areas and measuring the reflected light intensity for each small area;
Envelope value calculation step of calculating the envelope value by enveloping the reflected light intensity of the small area arranged along the envelope direction, with the application direction or the direction orthogonal to the application direction as the envelope direction,
Applying the application direction or the direction orthogonal to the application direction as the integration direction, integrating the envelope values arranged along the integration direction, and applying a portion where the integrated envelope value is larger or smaller than the surrounding integration value A non-uniformity extraction process for extracting as an area where there is a possibility of non-uniformity,
A coating unevenness inspection method comprising: In the method of inspecting the coating unevenness of the transparent resin film applied on the substrate, wherein the substrate has a lattice present in the lower layer of the transparent resin film,
A reflected light intensity measuring step of dividing the application direction of the transparent resin and a plurality of small areas by dividing the direction into the direction orthogonal thereto, irradiating light to these small areas and measuring the reflected light intensity for each small area;
A moving average value calculating step for calculating the moving average value by moving average the reflected light intensity of the small areas arranged along the moving average direction, with the application direction or the direction orthogonal to the application direction as the moving average direction;
Envelope value calculation step of calculating the envelope value by enveloping the reflected light intensity of the small area arranged along the envelope direction, with the application direction or the direction orthogonal to the application direction as the envelope direction,
A non-uniformity extraction step of extracting coating non-uniformity by a comparison operation between the moving average value and the envelope value,
A coating unevenness inspection method comprising:   2. The coating unevenness inspection method according to claim 1, wherein the envelope values calculated in the envelope value calculating step are an upper envelope value and a lower envelope value.   The comparison operation in the unevenness extraction step is a division of the upper envelope value and the moving average value, and a division of the lower envelope value and the moving average value. 3. The method for inspecting coating unevenness according to claim 2, wherein a possible area is extracted. The comparison operation in the unevenness extraction step is performed by dividing the upper envelope value and the moving average value, dividing the lower envelope value and the moving average value, and further multiplying each divided value to obtain an upper envelope. The method for inspecting unevenness in coating according to claim 2, wherein a region where there is a possibility of unevenness in coating is extracted from a difference between a ratio between a value and a lower envelope value and a threshold value. A program for inspecting coating unevenness from the reflected light intensity of the light irradiated to these small areas by dividing the transparent resin applied on the base material into a plurality of small areas by dividing the transparent resin in the coating direction and the direction perpendicular thereto. In this program, the base material has a lattice existing in the lower layer of the transparent resin film.
The envelope direction is a direction perpendicular to the application direction or the application direction, and the envelope value is calculated by enveloping the reflected light intensity of the small regions arranged along the envelope direction; and
Applying the application direction or the direction orthogonal to the application direction as the integration direction, integrating the envelope values arranged along the integration direction, and applying a portion where the integrated envelope value is larger or smaller than the surrounding integration value Extracting as an area with a possibility of unevenness;
A coating unevenness inspection program characterized by comprising: A program for inspecting coating unevenness from the reflected light intensity of the light irradiated to these small areas by dividing the transparent resin applied on the base material into a plurality of small areas by dividing the transparent resin in the coating direction and the direction perpendicular thereto. In this program, the base material has a lattice existing in the lower layer of the transparent resin film.
Applying a moving average of reflected light intensities of small areas arranged along the moving average direction, and calculating a moving average value thereof, with a coating direction or a direction perpendicular to the coating direction as a moving average direction;
The envelope direction is a direction perpendicular to the application direction or the application direction, and the envelope value is calculated by enveloping the reflected light intensity of the small regions arranged along the envelope direction; and
Extracting application unevenness by a comparison operation between the moving average value and the envelope value;
A coating unevenness inspection program characterized by comprising: