JPH05272949A - Method for evaluating glass plate surface - Google Patents

Abstract

PURPOSE:To quantitatively evaluate unevenness of polishing by estimating an intensity distribution of brightness fringes of optically reflected images generated by unevenness of a glass plate surface from direct measurement results of its shape by means of ray tracing, and by evaluating flatness of the glass plate with an amplitude value of the intensity distribution as an index. CONSTITUTION:An optical multi-line unevenness measuring device 12 is provided on a precise carrying mount 11 for carrying a glass polished plate 10. The measuring device 12 applies for example seven laser beams to the polished plate 10 with equal intervals, receives reflected light by a line sensor and generates analog unevenness information signals. This unevenness information signal is digitally converted by an A/D-converter 13 and input to a data processor 14. The data processor 14 calculates an intensity distribution of brightness fringes of an uneven reflected image by means of ray tracing, detects intensity difference in brightness of the intensity distribution by numeral processing and selects one with the largest difference in brightness to use it as a flatness index. Then it determines good and bad products to be displayed on a display 15 and also discriminates good products from bad ones via a controller 16 by a bad product discriminator 17.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a glass plate surface evaluation method, and more particularly to a glass plate surface evaluation method effective for inspecting the surface flatness of a liquid crystal glass substrate.

[0002]

2. Description of the Related Art A glass substrate for liquid crystal is a glass plate which requires a flatness of the glass surface in spite of its small thickness. Conventionally, such a highly flat glass plate has been manufactured by a float type manufacturing apparatus, but the flatness of the glass substrate for liquid crystal in recent years has reached a level far exceeding it, and therefore, a normal float type manufacturing apparatus is used. The produced glass plate is further polished by a polishing machine.

In the production of a highly flat glass plate by such polishing, a defect of uneven polishing sometimes occurs. The polishing unevenness includes large unevenness of 20 to 30 mm period and small unevenness of 3 to 8 mm period superimposed on the large unevenness. When a glass plate having uneven polishing is used for a liquid crystal display device, uneven color occurs due to uneven polishing of the above-described period. Polishing unevenness shows a state in which minute irregularities on the surface that existed on the surface of the glass plate before polishing remained after polishing, but in the polishing work, the degree of this polishing unevenness was quantitatively grasped, It is necessary to decide whether it will be discarded as defective.

Conventionally, as a method of evaluating the degree of unevenness of polishing, a method of irradiating light on a polishing surface and projecting the reflected light on a screen to functionally evaluate the pattern of bright and dark stripes that appeared is adopted. Has been. FIG. 2 shows this light reflection method. Light emitted from a light source, for example, a high-intensity lamp light source 2, is obliquely incident on a substrate glass 1 to be inspected, and light reflected by the glass surface is projected on a screen 3.
At this time, bright and dark stripes 4 appear on the screen 3 according to the shape of the glass surface, and the flatness of the substrate glass surface is evaluated by the intensity of the bright and dark stripes. That is, the minute irregularities on the surface of the glass plate act as a concave mirror or a convex mirror when forming a reflected image. When the glass surface is locally concave with respect to the outside, the reflected image corresponding to the position becomes a bright stripe, and conversely, when the glass surface is convex, a dark stripe is formed. In the conventional sensory evaluation, the intensity of the bright and dark stripes is used as an index of the flatness of the glass plate.

[0005]

This sensory evaluation method of flatness by the light reflection method makes it possible to easily evaluate a large glass plate, and its light and dark patterns are obtained when the substrate glass is assembled into a liquid crystal display device. Although it has the advantage of being able to deal well with the drawback of color unevenness of the liquid crystal of 2., it is a sensory evaluation, so it takes a lot of manpower to evaluate it because of individual differences, and technically it also prevents back reflection to prevent back reflection. It had a drawback that it was necessary to apply the above materials.

In view of such conventional problems, the present invention has
With the purpose of providing a method for evaluating polishing unevenness unattended, and for quantitatively evaluating polishing unevenness without impairing the correspondence between polishing unevenness accumulated by the conventional light reflection method and liquid crystal color unevenness. There is.

[0007]

In order to evaluate the flatness of the glass plate surface, the present invention directly measures the shape of the glass plate surface and uses the ray tracing method from the surface shape obtained by this measurement to It is characterized by predicting the intensity distribution of the bright and dark stripes of the light reflection image generated by the unevenness on the glass plate surface and evaluating the flatness of the glass plate surface using the amplitude value of the intensity distribution as an index.

Further, according to the present invention, in the prediction by the ray tracing method, the distance from the virtual light source to the virtual irradiation position of the virtual glass plate having the surface shape obtained by the direct measurement is calculated. The first step, the second step of calculating the position where the light reflected at the virtual irradiation position of the virtual glass plate is projected on the virtual screen, and the light intensity of the virtual light source in the first step The third step of calculating the light intensity at the virtual irradiation position by dividing by the square of the calculated distance and the virtual screen are divided into sections at equal intervals, and in each of the sections, the second step is performed. A fourth step of adding the light intensity calculated in the third step to a variable representing the ratio of the calculated projection position,
A fifth step of changing the virtual irradiation position and repeating the first to fourth steps to create a histogram by the fourth step, wherein the histogram is used for intensity distribution of bright and dark stripes of a light reflection image. It is characterized by

Further, according to the present invention, the direct measurement of the shape of the surface of the glass plate is performed along a plurality of parallel straight lines with a predetermined interval on the glass plate, and the plurality of surfaces obtained by this measurement are measured. The histogram is created for each of the shapes.

[0010]

In the conventional sensory evaluation by the light reflection method, the intensity of the bright and dark stripes of the reflected image is used as the index of the flatness of the glass plate.
The intensity of the bright and dark fringes can be predicted by the ray tracing method from the unevenness shape of the glass surface directly measured.

In the present invention, when predicting the intensity of bright and dark fringes by the ray tracing method, the intensity of the incident light on the virtual polishing plate is calculated as being equal to the intensity of the light projected on the virtual screen. It is not complicated and therefore does not take long to calculate. The amplitude value of the portion corresponding to the sharp bright and dark stripes present in the intensity distribution of the stripes of the reflection image obtained by calculation is detected, and this amplitude value is used as an index of the flatness of the glass plate surface. According to the glass plate surface evaluation method of the present invention, the same evaluation as the conventional sensory evaluation is possible, and the evaluation method of the present invention follows the principle of the conventional method and has no color unevenness of the liquid crystal cultivated in the past. It does not spoil the correspondence of.

[0012]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 3 shows a polishing plate determining apparatus in which the present invention is implemented. An optical multi-line unevenness measuring device 1 is mounted on a precision carrier 11 that carries a polishing plate 10 which is a glass substrate for liquid crystal.
Two are provided. An A / D converter 13 for converting its analog output into a digital output is connected to this measuring device, and this A / D converter 13 is connected to a data processing device 14. The data processing device 14 outputs information to a non-defective / defective product determination display 15 and a defective product sorter control device 16. The defective product sorter control device 16 supplies a control signal to the defective product sorter 17 provided on the precision carrier 11.

Optical multi-line unevenness measuring device 12, A /
The D converter 13, the computer 14, and the non-defective / defective product judgment display 15 are related to the glass plate surface evaluation method of the present invention. As the optical multi-line unevenness measuring device 12, for example, a commercially available optical displacement meter with high resolution is used. This optical displacement meter uses a laser as a light source, irradiates the polishing plate 10 with laser light obliquely, receives the reflected light with a line sensor, and measures the uneven shape of the entire surface of the polishing plate. The data processing device 14 performs a ray tracing method on the obtained concavo-convex shape to predict a reflection image on the glass plate surface. The intensity of the bright and dark fringes of the reflection image obtained by this prediction is detected and used as an index of the flatness of the glass plate surface. Based on the index, the data processing device 14 determines whether the product is a non-defective product or a defective product, displays the determination result on the non-defective product / defective product determination indicator 15, and sends the determination result to the defective product sorter control device 16. Control device 1
In 6, the defective product sorting machine 17 is controlled according to the determination result, and the polishing plate 10 sent on the precision carrier 11 is sorted into good products and defective products.

The optical multi-line unevenness measuring device 12 irradiates a polishing plate 10 of, for example, 400 × 400 mm with seven laser beams (for example, a diameter of 7 μm) 9 at equal intervals as shown in FIG. The reflected light is received by the line sensor and an analog unevenness information signal is generated. This unevenness information signal is converted into digital unevenness data by the A / D converter 13, and is taken into the memory of the data processing device 14. The unevenness data includes, for example, numerical data of 2048 points for one laser beam.

The data processor 14 calculates the intensity distribution of the bright and dark fringes of the reflected image by the ray tracing method which will be described later based on the numerical data of each of the seven laser beams. Then, the calculated brightness difference (amplitude value) of the intensity distribution is detected by numerical processing. Then, the one having the largest intensity difference between bright and dark is selected, and the intensity difference is used as an index of flatness.

Next, the calculation of the intensity distribution of the bright and dark fringes of the reflection image by the ray tracing method will be described with reference to the flowchart of FIG. In the following, the calculation from the unevenness data for one of the seven laser beams will be described.
Others are the same.

When the measurement direction of the polishing plate by the unevenness measuring device 12 is x and the thickness direction of the polishing plate is y, the unevenness data f
(X) is shown in FIG. Then, the x-coordinate of the left end of the polishing plate is initially set to X ₀ (usually X ₀ = 0.0) and the x-coordinate of the right end is initially set to X _E, and the unit width of the calculation is ΔX (for example, 0.001 m).
m) is initialized (step S1).

When the coordinate of an arbitrary point on the x-axis of FIG. 5 is X, X = X ₀ is set (step S2), and the differential coefficient f '(X) at the coordinate X is calculated (step S3). ). FIG. 6 shows an enlarged view around the coordinate X. The angle formed by the tangent line of the irregularity data f (x) at the coordinate X to the x coordinate axis is θ _A, and θ _A is calculated by tan θ _A = f ′ (X) (step S4).

FIG. 7 is a conceptual diagram when the data processor 14 calculates the intensity distribution of the bright and dark stripes of the reflected image. xy
Virtual polishing plate 10 'with the origin (0, 0) of the coordinate axis as the left end
Are arranged in the x-axis direction. Further, in the xy coordinate system of FIG. 7, the coordinates of the position of the virtual light source 2 ′ are set to P
(X _L , Y _L ), and the virtual screen 3'is assumed to be provided at the x coordinate position X _R in parallel with the y coordinate axis. It is assumed that the light from the virtual light source 2'is reflected at the point Q (x coordinate position X) on the virtual polishing plate 10 'and projected on the virtual screen 3'. The angle between the incident light and the x coordinate axis is θ ₁ , and the reflected light is x
The angle with the axis is θ ₂ . When the point Q portion of the virtual polishing plate 10 'is flat along the x axis, the angle formed by the reflected light and the x coordinate axis is θ ₁ , and θ ₂ −θ ₁ = 2θ _A. θ _A is
It is the angle θ _A described in FIG.

The angle θ ₁ is calculated from tan θ ₁ = Y _L / (X _L + X) (step S5). From θ _A obtained in step S4 and θ ₁ obtained in step S5, the angle θ ₂ is calculated by θ ₂ = θ ₁ + 2θ _A (step S6).

[0022] If the projection point of the reflected light of the virtual screen 3 'above was D _X, the y coordinate position Y _DX of D _X, is calculated by _{tanθ 2 = Y DX / (X} R -X) ( step S7 ). Then, the distance T _X between the position P of the virtual light source 2'and the incident position Q on the virtual polishing plate 10 'is calculated by T _X ² = Y _L ² + ( _XL + X) ² (step S8). When the light intensity of the virtual light source 2'is I ₀ , the incident point Q on the virtual polishing plate 10 '
The light intensity I _{X at} is calculated by I _X = I ₀ / (T _X ) ² (step S9). The intensity of the light projected on the point D _X on the virtual screen 3'is equal to the light intensity I _X on the point Q on the virtual polishing plate. Actually, the light intensity of the point D _X on the virtual screen is (distance PQ
+ Inversely proportional to the square of the distance QD _X ), but in that case Δ
X is not evenly spaced, and its calculation requires a lot of time. To avoid this, set ΔX at equal intervals as described above,
Instead, the light reflected at the point Q on the virtual polishing plate reaches the point D _X on the virtual screen as it is without being attenuated. In this way, even if the intensity distribution of the bright and dark fringes of the reflection image is calculated, the accuracy of evaluation is hardly affected.

The virtual screen 3'is divided at equal intervals (for example, 1.0 mm intervals) as shown in FIG. 8 and variables E (11), E (1
2), E (13) ... And the light intensity _IX of the projected light is added to this variable (step S10). For example Y _DX =-
Assuming 12.3, the light intensity I _X is added to the variable E (12). This process is for creating a histogram representing the intensity distribution of the bright and dark stripes of the reflected image, as will be described later.

When the above calculation is completed, ΔX is added to X (step S11), the process returns to step S3, and the same process is repeated. If X> X _{E in} step S12, the process ends.

Through the above processing, a histogram is created in step S10, which represents the intensity distribution of the calculated bright and dark fringes of the reflected image.

FIG. 9 (a) shows the surface roughness of the unpolished plate as shown in FIG.
(B) shows the intensity distribution of the bright and dark fringes of the reflection image calculated by the method of this example for this unpolished plate. According to FIG. 9A, it can be seen that the maximum unevenness is about 0.1 μm within the measurement length of 200 mm. In the calculated intensity distribution of FIG. 9B, the amplitude value at the position A is maximum,
It can be seen that sharp bright and dark stripes exist at this position.

FIG. 10 (a) shows the surface irregularities of the polishing plate which is a non-defective product, and FIG. 10 (b) shows the intensity distribution of the bright and dark stripes of the reflection image calculated for this non-defective product. Since this polishing plate is a good product, it can be seen that the amplitude value is small over the entire surface.

The evaluation of non-defective products and defective products of the polishing plate is performed by using the calculated amplitude value of the light intensity distribution as an index of flatness. For example, whether or not the maximum amplitude value exceeds the reference value, or It is possible to evaluate non-defective products and defective products by the number of amplitude values that exceed. How to select the evaluation judgment is not the gist of the present invention.

[0029]

According to the present invention, the flatness of a glass plate can be quantitatively and automatically determined without impairing the evaluation criteria accumulated by the conventional sensual evaluation method of light and dark stripe patterns by the light reflection method. It is possible to evaluate.

[Brief description of drawings]

FIG. 1 is a flowchart showing an embodiment of the present invention.

FIG. 2 is a diagram illustrating a conventional light reflection method.

FIG. 3 is a diagram showing a polishing plate determination device for carrying out the method of the present invention.

FIG. 4 is a diagram showing the number of lines measured by an optical multi-line unevenness measuring device.

FIG. 5 is a diagram showing unevenness data f (x).

FIG. 6 is an enlarged view of a coordinate X portion.

FIG. 7 is a diagram for explaining the concept of calculation by the ray tracing method.

FIG. 8 is a diagram for explaining how to create a histogram.

FIG. 9 is a diagram showing surface irregularities of an unpolished plate and a corresponding intensity distribution by calculation.

FIG. 10 is a diagram showing surface irregularities of a polishing plate and corresponding intensity distributions by calculation.

[Explanation of symbols]

 10 Polishing Plate 11 Precision Carrying Platform 12 Optical Multi-Line Concavo-Convex Measuring Device 13 A / D Converter 14 Data Processing Device 15 Good / Defective Product Judgment Indicator 16 Control Device 17 Defective Product Sorting Machine

Claims (3)

[Claims] 1. When evaluating the flatness of a glass plate surface, the shape of the glass plate surface is directly measured, and the surface shape obtained by this measurement is generated by unevenness on the surface of the glass plate by a ray tracing method. A glass plate surface evaluation method characterized by predicting the intensity distribution of bright and dark stripes of a light reflection image, and evaluating the flatness of the glass plate surface using the amplitude value of the intensity distribution as an index. 2. The prediction by the ray tracing method includes a first step of calculating a distance from a virtual light source to a virtual irradiation position of a virtual glass plate having a surface shape obtained by the direct measurement. A second step of calculating a position where the light reflected at the virtual irradiation position of the virtual glass plate is projected on a virtual screen, and a light intensity of the virtual light source of the distance calculated in the first step of 2 The third step of calculating the light intensity at the virtual irradiation position by dividing by the power, and dividing the virtual screen into sections at equal intervals, and the projection position calculated in the second step for each section Is added to the variable representing the ratio of the existence of the light intensity calculated in the third step, and the virtual irradiation position is made different, and the first to fourth steps are repeated, In 4 steps Fifth and a step, the glass plate surface evaluation method according to claim 1, characterized in that the intensity distribution of light and dark stripes of light reflection image the histogram to create that histogram. 3. The direct measurement of the shape of the surface of the glass plate is performed along a plurality of parallel straight lines with a predetermined interval on the glass plate, and for each of the plurality of surface shapes obtained by this measurement, The glass plate surface evaluation method according to claim 2, wherein the histogram is created.