JP2008064759A - Device and method for optically detecting surface error of substrate - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device allowing easy detection of a fine scratch. <P>SOLUTION: This device optically detects a surface error of a substrate 10, especially a fine scratch. The device comprises an illumination 15 for illuminating a detection region of the substrate 10 surface, and a detector 18 for detecting the light radiated from the illumination directed to the detection region. The illumination 15 and detector 18 are arranged so that the illumination light 20' scattered by the surface error in the detection region reaches the detector 18 at least partially and the illumination light 20" scattered in the detection region having no substrate surface error does not reach the detector 18. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an apparatus for optically detecting a substrate surface error, particularly a fine scratch.

  In general, the detection of surface errors, in particular scratches, is particularly important in the production of glass substrates for forming flat displays, in particular flat displays based on TFT technology. This is because here a particularly high demand is placed on the quality of the surface properties of the glass substrate.

  Glass substrates used to form TFT flat displays typically today have a size of about 2000 mm × 2000 mm at a thickness of 0.4 mm to 1.1 mm. When manufacturing this glass substrate, the surface of the substrate is polished. In that case, a small error, such as a scratch, may occur on the surface of the substrate, or a surface error that has already occurred before that may not be completely removed by polishing.

  There are surface inspection systems that allow the detection of glass errors resulting from glass manufacturing, such as local deformations or bubbles. However, the detection of relatively small surface errors, especially fine scratches, has been shown to be a problem so far. Here, the fine scratch is a scratch having a depth of a nanometer region and a width of a region of several tens to several hundreds of nm in an arbitrary length.

  Conventionally, detection of scratches on a glass substrate has been performed using human observation. The detection results depend on the incidence of light to a significant extent because the formation of scratches is substantially two-dimensional, so a person who is requested to detect scratches tilts the glass substrate to be inspected with respect to the light source It must be possible. However, this is difficult for thin substrates having a size of an area of several square meters.

  It is also known to detect scratches on a glass substrate using scattered light in a bright field arrangement. However, based on diffraction limitations, this method can only detect relatively large scratches and not fine scratches having the dimensions described above. Furthermore, the glass substrate to be inspected is rotated relative to the illumination and / or relative to the photosensitive detector of the detection device, as described above, because the scratches are formed substantially in two dimensions. This type of detection device is not suitable for use in a continuous process because it must be tilted.

  An object of the present invention is to provide an apparatus that allows easier detection of fine scratches.

  In order to solve this problem, an apparatus having the features of claim 1 is provided.

  The device according to the invention comprises an illumination that illuminates a detection area on the substrate surface and a detector that detects light emitted from the illumination that is directed to the detection area.

  The illumination and the detector are as follows with respect to each other, i.e., the detection light that is scattered at the surface error that is within the detection area and at least partially reaches the detector and is free of errors on the substrate surface The illumination light scattered within is arranged so that it does not reach the detector. In other words, therefore, the arrangement according to the invention of the illumination and detector is a dark field arrangement which allows detection of a much smaller structure compared to the bright field arrangement. That is, a fine scratch having, for example, a depth of 10 nm and a width of 100 nm is detected by the apparatus according to the invention.

  The device according to the invention further comprises an optical aperture magnifying means, the aperture magnifying means being arranged between the illumination and the substrate surface, and being emitted from the illumination by the aperture magnifying means. The aperture of light can be expanded. Aperture magnification means are used to maximize the area of the angle at which the light beam emitted from the illumination sees in the projection onto the substrate surface and abuts against the scratch. By appropriately forming the aperture magnifying means and appropriately positioning the aperture magnifying means relative to the substrate surface, for example, an angular region of approximately 180 ° of the light beam can be achieved within the projection onto the substrate surface. Is possible.

  This expansion of the light emitted from the illumination makes error detection according to the invention substantially independent of the position or orientation of the scratch. In this way, fine scratches can be reliably detected without rotating and / or tilting the substrate to be inspected with respect to the illumination or detector. Thereby, according to the invention, scratch detection is not only greatly simplified, but can also be carried out in a continuous process.

  Note that the device according to the invention is described here with the detection of fine scratches in particular, but is not limited to the detection of fine scratches, and in principle other, eg three-dimensional It is also suitable for detecting surface errors.

  Furthermore, the apparatus according to the present invention is not limited to inspection of glass substrates, but is also used for surface inspection of semiconductor substrates, such as silicon wafers, for example.

  Preferred forms of the invention will become apparent from the dependent claims, the description and the drawings.

  According to certain embodiments, the illumination and detector are located on the same side of the substrate surface. This type of arrangement is also referred to as a reflected light-dark field arrangement. This allows the detection of surface errors regardless of whether the substrate is transparent, i.e. whether it transmits light emitted from the illumination.

  According to another embodiment, the illumination is linear illumination. The light emitted from the linear illumination defines a scanning plane that, when abutted against the substrate surface, results in an elongated, in particular at least approximately linear detection region. This allows for quick and accurate scanning of large area substrates.

  Preferably, the illumination is oriented transversely and in particular at right angles to the substrate feed direction. This enables a particularly effective inspection of the substrate moving through the illumination. The illumination can be dimensioned so that the detection area extends across the entire width of the substrate. Alternatively, a plurality of side-by-side lighting can be provided for this purpose. In both cases, if the detector is configured accordingly, a single pass of the substrate through the detector is sufficient for inspection of the substrate surface. This detection device is therefore particularly suitable for detecting surface errors in a continuous process.

  Preferably, the aperture magnifying means is only effective at least approximately in the direction of the longitudinal extent of the illumination, i.e. in the light scanning plane. In this way, it is ensured that the linearity of the detection area is not impaired by the aperture enlargement means.

  According to another embodiment, the aperture magnifying means comprises a beam splitter and a lenticular lens arranged especially in the rear stage of the beam splitter in the light propagation direction. This kind of arrangement of beam splitters and lenticular lenses results in a particularly large broadening of the light emitted from the illumination, ie a particularly significant enlargement of the aperture. Furthermore, this type of aperture enlargement means provides a sufficiently high light intensity for the detection of fine scratches over the entire angular region.

  The beam splitter can have a plurality of prisms arranged one after the other (as viewed in the direction of the longitudinal extent of the illumination). Similarly, a lenticular lens can also have a plurality of cylindrical lenses arranged one after the other (as viewed in the direction of the longitudinal extent of the illumination). Thus, both beam splitters and lenticular lenses are Fresnel optical components.

  According to another embodiment, the illumination has a cross-section converter, by which the cross-section of the light emitted from the light source of the illumination can be converted into at least a substantially linear cross-section. is there. With a cross-section converter, it is possible to use a cold light source, such as a halogen lamp, arc lamp or halogen metal vapor deposition lamp, to generate a linear detection area.

  A simple form of cross-section converter has, for example, a plurality of light guiding fibers, which are bundled in the region of the light source and are connected to each other in the direction of the light exit aperture of the illumination. It extends away.

  According to another embodiment, at least one beam trap is provided for capturing illumination light that is not scattered in the direction of the detector. The light that is not scattered in the direction of the detector is, for example, light reflected by the substrate surface without error or light transmitted through the substrate. Capturing this light in the beam trap prevents the light from reaching the detector undesirably and distorting the detection results.

  According to another embodiment, the detector has at least one camera. This camera is, for example, a CCD camera.

  Preferably, the detector comprises a plurality of cameras arranged one after the other (as viewed in the direction of the longitudinal extent of the illumination). Preferably, the number of cameras is adapted to the length of the detection area. The use of multiple cameras supports rapid inspection of large area substrates.

  One or each camera can in particular be a line camera oriented in parallel to an elongated, preferably at least approximately linear detection region. For example, in order to enable efficient inspection of a large area substrate in a continuous process, one or each line camera is oriented transversely to the substrate feed direction depending on the linear detection area. be able to.

  Preferably, one or each camera is a TDI camera (TDI “time delay and integration”) with multiple rows of photosensitive elements, in which case the rows are in phase (as viewed in the substrate feed direction). Are arranged. In this type of TDI camera, the camera signal is integrated with the substrate feed clock, thereby achieving particularly high sensitivity, which contributes to reliable detection of fine scratches in the dark field.

  Another subject of the present invention is also a method having the features of claim 18. With the method according to the invention, the advantages mentioned above are likewise achieved.

  The present invention will now be described using preferred embodiments purely by way of example with reference to the accompanying drawings.

  FIG. 1 shows an apparatus according to the invention for optically detecting surface errors of a substrate 10, in particular fine scratches. In this embodiment, the substrate 10 is a glass substrate having a thickness in the range of 0.4 mm to 1.1 mm provided for use in a TFT flat display. However, in principle, the substrate 10 is not limited to a glass substrate, and may be a semiconductor substrate such as a silicon wafer.

  The substrate 10 is held on a transfer device not shown in order to enable error detection in a continuous process and can be moved in the feed direction 12 relative to the detection device by this transfer device.

  The detection device comprises an illumination with a cross-section converter 15 for illuminating the detection area 16 of the substrate 10 and a detector 18 for detecting the light 20 emitted from the illumination directed to the detection area 16. And have.

  Illumination with a cross-sectional transducer 15 and detector 18 are illuminated light 20 ′ scattered on the same side of the substrate 10 with respect to each other as follows, ie with fine scratches in the detection region 16. So that the illumination light 20 ″ with the cross-section converter 15, which is at least partly reaching the detector 18 and scattered in the error-free detection region 16 on the substrate surface, cannot reach the detector 18. Accordingly, the arrangement of the illumination with the cross-section converter 15 and the detector 18 is, in other words, a reflected light-dark field arrangement.

  The optical axis of the detector 18 is oriented perpendicular to the surface of the substrate 10. In order to obtain a detection result that is optimal and as independent as possible from the orientation of the fine scratches to be detected, the irradiation of the detection region 16 should be carried out at as steep angles as will be described in detail below. In this embodiment, the optical axis of the illumination with cross-sectional transducer 15 forms an angle of about 9 ° with the optical axis of detector 18.

  The illumination has a cold light source such as a halogen lamp, an arc lamp or a halogen-metal vapor deposition lamp, and a cross-sectional converter 15 is connected to the subsequent stage as viewed in the light propagation direction.

  The cross-section converter 15 has fibers for guiding a plurality of lights, and these fibers form a light source of the cross-section converter 15 in order to form a point beam on the light exit side of the cross-section converter 15. The light beams are substantially bundled into a circular light beam on the side facing, and extend so as to spread in a fan shape in a plane as seen in the light propagation direction, that is, a so-called scanning plane of emitted light.

  A cylindrical lens 22 is connected to the rear stage of the cross-section converter 15 as viewed in the light propagation direction, the longitudinal axis of the cylindrical lens is in the scanning plane, and is parallel to the surface of the substrate 10 and in the feed direction. 1 extends at right angles to 12 and thus in FIG. 1 at right angles to the drawing plane. A cylindrical lens 22 has the shape of a line in which the detection region 16 extends at least approximately perpendicular to the feed direction 12 and thus perpendicular to the drawing plane in FIG. To focus. Thus, the illumination is ultimately line illumination.

  An aperture magnifying means 24 is connected to the rear stage of the cylindrical lens 22 when viewed in the light propagation direction, and the aperture magnifying means transmits the aperture of the light 20 emitted from the illumination having the cross-section converter 15 within the scanning plane. Used to enlarge as much as possible.

  The aperture enlarging means 24 has a beam splitter 26 and a lenticular lens 28 (FIG. 2) that are arranged one after the other in the light propagation direction. The beam splitter 26 and the lenticular lens 28 each have a rod-like basic shape, and each extend parallel to the cylindrical lens 22 and thus perpendicular to the drawing plane in FIG.

  As shown in FIG. 2 a, the beam splitter 26 has a large number of prisms 30 arranged side by side in the longitudinal direction of the beam splitter 26, and these prisms are light beams that abut on the beam splitter 26. In order to split the beam 20 into two light beams 20a and 20b extending away from each other in the scanning plane, each has a triangular cross section in the scanning plane.

  As is clear from FIG. 2 b, the lenticular lens 28 has a large number of cylindrical lenses 32 arranged side by side in the longitudinal direction of the lenticular lens 28, and these cylindrical lenses are incident on the lenticular lens 28. In order to convert the collimated light beam 20 into a diverging light beam 20c within the scanning plane, each has a partial circular cross section within the scanning plane.

  In FIG. 3, the action of the beam splitter 26 and the lenticular lens 28 is shown according to the angle in the scanning plane using the intensity of light passing through them. The light intensity described in that case is in the main radiation direction of the initial light beam emanating from the illumination with the cross-section converter 15 and not affected by the beam splitter 26 or lenticular lens 28, ie 0 °. Is normalized with respect to the maximum intensity measured at a radiation angle of.

  The intensity distribution of the initial light beam emitted from the illumination having the cross-section converter 15 and not affected by the beam splitter 26 or the lenticular lens 28 is shown by a solid curve (a). As is apparent from curve (a), the light emitted from the illumination having the cross-section converter 15 has a relatively slight divergence. More precisely, the initial light beam has a measurable intensity only in the region of about −40 ° to about + 40 °, in which case the normalized intensity is −20 ° to +20 It takes a value greater than 0.2 for the first time in an angular region of °.

  A dotted curve (b) shows the intensity distribution of the light after passing through the beam splitter 26. As is apparent from this curve, intensity maxima are formed with normalized intensities of about 0.55, respectively, under the radiation angles of about −25 ° and about + 25 °, whereas the intensity is in the main radial direction. I.e. about 0.15 under a radiation angle of about 0 °. Thus, the beam splitter 26 splits the incident light beam 20 into two partial beams 20a and 20b, as already described with reference to FIG. 2a.

  A broken curve (c) indicates the light intensity of the light after passing through the lenticular lens 28. As is apparent from curve (c), lenticular lens 28 provides a clear broadening of the intensity distribution as compared to the intensity distribution of the initial light beam, as shown by the solid curve (a). In other words, the lenticular lens 28 provides a widening of the incident light beam (see also FIG. 2b).

  An alternate long and short dash line curve (d) represents an intensity distribution of light that has passed through the combination of the beam splitter 26 and the lenticular lens 28. As is evident from curve (d), the combination of beam splitter 26 and lenticular lens 28 results in the light beam passing through this combination having a significant intensity over an angular region greater than −60 ° to + 60 °. .

  In other words, therefore, the combination of the beam splitter 26 and the lenticular lens 28 results in a beam broadening that is much larger than the beam widening that can be obtained by the lenticular lens 28 itself. Furthermore, the combination of beam splitter 26 and lenticular lens 28 forms a much more homogeneous intensity distribution over the -60 ° to + 60 ° angular region than can be achieved by beam splitter 26 or lenticular lens 28, respectively. . Thus, the combination of beam splitter 26 and lenticular lens 28 allows for homogeneous illumination of detection region 16 having an aperture greater than 120 °.

  Therefore, by appropriately forming the aperture magnifying means 24, that is, by appropriately forming and arranging the beam splitter 26 and the lenticular lens 28, the detection region 16 is finally made homogeneous and within the scanning plane. With an aperture of the following size of the emitted light 20, i.e. when the emitted light 20 is viewed in projection onto the substrate surface, an angle of about 180 ° on the surface error, e.g. a fine scratch Irradiation is possible with an aperture that is sized to abut in the region.

  The angular area achievable in the projection onto the substrate surface in that case depends proportionally on the steepness of the light incidence, i.e. the angular area depends on the optical axis or scanning plane of the illumination with the cross-section converter 15. The larger the angle to be formed, the larger it becomes. Thus, as already explained, the angle formed by the optical axis of the illumination with the transverse detector 15 and the optical axis of the detector 18 should be as small as possible, and therefore within the dark field arrangement of the illumination and detector. It should be small so that it can be realized.

  The angular region or aperture of the light 20 that abuts on the substrate surface is substantially, i.e., approximately 180 ° in projection onto the substrate surface, so that scratches characterized by its almost two-dimensional formation, particularly fine Scratches can also be detected almost regardless of their orientation. In particular, not only fine scratches extending in the transverse direction with respect to the feed direction 12 but also fine scratches substantially oriented in the feed direction 12 can be detected reliably.

  As already explained, since the optical axis of the detector 18 is oriented perpendicular to the surface of the substrate 10, the illumination light 20 'is scattered by surface errors of the substrate, for example fine scratches. Only when this is done can the detector 18 be reached.

  As is known per se, a mask 34 and an objective lens 36 are provided in front of the detector 18 when viewed in the light propagation direction.

  The detector 18 itself has a plurality of CCD linear cameras, each oriented parallel to the detection region 16 and thus perpendicular to the drawing plane in FIG. The linear cameras are arranged side by side when viewed in the lateral direction with respect to the feeding direction 12, and in this case, the length and number of the linear cameras are selected so that the entire length of the detection region can be monitored. The cameras each have a scan length of, for example, 100 mm, and 5 cameras can be associated with each illumination.

  The linear camera is a so-called TDI camera (TDI “time delay and integration”) having a plurality of rows of photosensitive elements. In this case, the rows are arranged one after the other in the feed direction 12. The individual row signals of each camera are integrated with the feed clock of the substrate 10, thereby obtaining increased camera sensitivity.

  Based on the increased sensitivity of the camera and the low intensity of the scattered light signal of fine scratches, in order to ensure sufficient contrast to the background for detection of fine scratches, The detection device must be sealed against ambient and scattered light. For this purpose, the detection device comprises a first beam trap 38 that removes the light 20 "reflected from the surface of the substrate 10 and a second beam trap that removes the light 20" transmitted through the substrate 10. 40. Therefore, neither the light 20 ′ reflected by the substrate surface nor the light 20 ″ transmitted through the substrate 10 by the beam traps 38, 40 can reach the imaging optical system and thus the detector 18 as scattered light.

It is the structure explanatory view which showed typically the apparatus based on this invention. It is explanatory drawing which shows the function (a) of the beam splitter of the apparatus of FIG. 1, and the function (b) of a lenticular lens. The intensity of light radiated from the illumination of the apparatus of FIG. 1 is expressed by the radiation angle (a) when exiting from the illumination, the radiation angle after passing through the beam splitter (b), and the radiation angle after passing through the lenticular lens ( It is the characteristic view shown as c) and the radiation angle (d) after passing through the combination of the beam splitter and the lenticular lens.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Board | substrate 12 Feed direction 15 Cross-section converter 16 Detection area 18 Detector 20 Light 22 Cylindrical lens 24 Aperture expansion means 26 Beam splitter 28 Lenticular lens 30 Prism 32 Cylindrical lens 34 Mask 36 Objective lens 38 Beam trap 40 Beam trap

Claims (18)

A device for optically detecting surface errors of the substrate (10), in particular fine scratches,
Illumination for illuminating the detection area (16) of the substrate surface;
A detector (18) directed to the detection region (16) for detecting light (20) emitted from the illumination;
Have
Illumination light (20 ′) is at least partially reaching the detector (18), as the illumination and the detector (18) are scattered with respect to each other as follows: Then, the illumination light (20 ″) scattered in the error-free detection area of the substrate is arranged so that it does not reach the detector (18), and optically between the illumination and the substrate surface. Aperture enlargement means (24) is arranged, and the aperture of the light (20) emitted from the illumination can be enlarged by the aperture enlargement means.
A device that optically detects substrate surface errors, especially fine scratches.   2. Device according to claim 1, characterized in that the illumination and the detector (18) are arranged on the same side of the substrate (10).   Device according to claim 1 or 2, characterized in that the illumination is linear illumination.   Device according to claim 3, characterized in that the illumination is oriented transversely and in particular perpendicularly to the feed direction (12) of the substrate (10).   Device according to at least one of the preceding claims, characterized in that the aperture expansion means (24) are effective only at least approximately in the direction of the longitudinal extent of the illumination.   6. A device according to at least one of the preceding claims, characterized in that the aperture expansion means (24) comprises at least one beam splitter (26).   7. Device according to at least one of the preceding claims, characterized in that the aperture magnifying means (24) comprises at least one lenticular lens (28).   The aperture magnifying means (24) has a beam splitter (26) and a lenticular lens (28) arranged particularly in the rear stage of the beam splitter (26) when viewed in the light propagation direction. An apparatus according to at least one of claims 1 to 7.   9. The beam splitter (26) according to claim 6 or 8, characterized in that it has a plurality of prisms (30) arranged one after the other when viewed in the direction of the longitudinal extent of the illumination. apparatus.   The lenticular lens (28) has a plurality of cylindrical lenses (32) arranged one after the other when viewed in the longitudinal direction of illumination. Equipment.   The illumination has a cross-section converter (15), by means of which the cross-section of the light emitted from the light source of the illumination can be converted into at least a substantially linear cross-section. An apparatus according to at least one of claims 1 to 10.   The cross-section converter (15) has a plurality of fibers for guiding light, said fibers being bundled in the region of the light source and extending away from each other in the direction of the light exit opening of the illumination. The apparatus according to claim 11, wherein   13. At least one beam trap (38, 40) is provided for capturing illumination light (20 ') that is not scattered in the direction of the detector (18). An apparatus according to at least one.   14. Apparatus according to at least one of the preceding claims, characterized in that the detector (18) comprises at least one camera.   15. The detector (18) according to at least one of the preceding claims, characterized in that the detector (18) has a plurality of cameras arranged one after the other when viewed in the direction of the longitudinal extension of the illumination. Equipment.   16. Device according to claim 14 or 15, characterized in that one camera or each camera is a linear camera, in particular oriented parallel to the elongated detection area (16).   One or each camera is a TDI camera having a plurality of rows of photosensitive elements, and the rows are arranged one after the other when viewed in the feed direction (12) of the substrate (10). An apparatus according to any one of claims 14 to 16. A method for optically detecting a surface error of a substrate (10), in particular a fine scratch, comprising the following steps:
Illumination is used to illuminate the detection area (16) of the substrate surface, and using a detector (18) directed to the detection area (16), light (20) emitted from the illumination is detected,
Illumination light (20) is scattered at least partially with respect to each other in the following manner, i.e. with a surface error in the detection region (16). And the illumination light (20 ″) scattered in the detection area (16) free of errors on the substrate surface is arranged so that it does not reach the detector (18), and the illumination and the substrate surface An optical aperture magnifying means (24) is disposed between and the aperture of the light (20) emitted from the illumination is magnified by the aperture magnifying means.
A method for optically detecting substrate surface errors.

Images