JP2009054649A - Identification symbol reading method and apparatus for semiconductor wafer - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide reading method and apparatus capable of reading an identification symbol even when the surface state changes because wirings, etc. are formed near the identification symbol marked on a semiconductor wafer. <P>SOLUTION: This identification symbol reading method concerns reading the identification symbol 12 consisting of an arrangement of a plurality of depressions marked on the semiconductor wafer 10 during intervals between a plurality of steps for manufacturing a semiconductor device on the semiconductor wafer by irradiating the depressions and the surroundings thereof with light 58. It comprises the first step for setting up an angle which the traveling direction of light 58 forms with the surface of the semiconductor wafer 10 at the position of the identification symbol so that the depressions can be observed more brightly than the surroundings, and the second step for reading the identification symbol after the first step. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a method for reading an identification symbol of a semiconductor wafer and an apparatus therefor, and more particularly to an identification symbol reading method that makes it possible to read an identification symbol that has become difficult to read due to a change in the state of the surface of the semiconductor wafer in the manufacturing process of the semiconductor device. And an apparatus for the same.

  A semiconductor device is manufactured by simultaneously processing a large number of semiconductor wafers on a manufacturing line composed of various manufacturing apparatuses.

  In such a production line, a different identification number (ID) is marked for each semiconductor wafer. By this identification symbol, processing history information (history of processing performed) and quality information (characteristics of manufactured semiconductor elements and measurement conditions thereof) of the semiconductor wafer are managed.

  In addition, the identification number is also used for replacement work in which the semiconductor wafers stored in one carrier are transferred to another carrier (for example, carriers of different materials) at a time. Furthermore, the identification number is also used for the rearrangement work for rearranging the semiconductor wafers in the same carrier.

  Thus, the identification symbol is indispensable for the manufacture of the semiconductor device.

  By the way, in order to mark the identification symbol on the semiconductor wafer, for example, the semiconductor wafer is irradiated with a laser beam to form fine depressions (dots). The identification symbol is formed by arranging such depressions so as to form, for example, characters or barcodes.

In order to read the identification symbol formed in this way, the identification symbol irradiated with light is photographed by a CCD camera or the like, and the obtained image is processed to extract the identification symbol (Patent Documents 1 and 2).
JP-A-6-112300 JP 7-307256 A

  However, if the state of the surface of the semiconductor wafer changes through various manufacturing processes, it becomes difficult to read the identification symbol and misreading occurs.

  The inventor reads an identification symbol marked on a semiconductor wafer by using two kinds of reading devices each having an illumination unit having a different structure.

  However, regardless of which device is used, the identification symbol is misread while passing through various manufacturing processes. Therefore, the mechanism of misreading will be described for both devices.

(1) Identification symbol reader with epi-illumination unit (Related technology 1)
(A) Configuration FIG. 19 is a schematic view of one identification symbol reading device 2 used by the present inventor as viewed from the side. This reading device is a device that reads the identification symbol 12 by irradiating the semiconductor wafer 10 with light generated by the epi-illumination unit 8.

  The identification symbol reading device 2 is used between a plurality of processes for manufacturing a semiconductor device on the semiconductor wafer 10. Then, the identification symbol reader 2 reads the identification symbol 12 that is marked on the semiconductor wafer 10 and includes an array of a plurality of recesses by irradiating the incident light 14 on the recesses and the periphery thereof.

  The configuration of the identification symbol reader 2 is as follows.

  First, the identification symbol reading device 2 includes an illumination 4 and a half mirror 6, and includes a plurality of depressions constituting the identification symbol 12 and an epi-illumination unit 8 that irradiates the epi-illumination light 14 on the periphery thereof.

  In addition, the identification symbol reading device includes a camera 16 that captures an image of an area composed of the depression and its periphery.

Further, the identification symbol reader 2 includes an image processing unit 18 that processes an image captured by the camera 16 and extracts the identification symbol 12.
(B) Reading Operation Reading of the identification symbol 12 using the apparatus shown in FIG. 19 is performed as follows.

  First, the illumination 4 is turned on to irradiate incident light on the depressions forming the identification symbol 12 and its surroundings.

  Next, the identification symbol 12 is photographed by the CCD camera 16 and the image photographed by the CCD camera 16 is processed to extract the identification symbol 12. That is, the identification symbol 12 is read. The incident light refers to light that is irradiated onto the semiconductor wafer 10 from directly above.

  FIG. 20 is a diagram illustrating a state in which the incident light 14 is irradiated to the depression 20 constituting the identification symbol 12 and the periphery 26 thereof. Most of the incident light 14 applied to the periphery 26 of the depression 20 is specularly reflected and becomes specularly reflected light 22. As a result, most of the reflected light 22 enters the CCD camera 16 disposed immediately above the semiconductor wafer 10. Here, regular reflection refers to reflection in which incident parallel rays become parallel rays after reflection.

  On the other hand, the incident light 14 applied to the depression 20 is reflected (diffusely reflected) in various directions by the curved surface constituting the surface of the depression 20 or the unevenness formed there, and the reflection direction is not constant. For this reason, only a part of the incident light 14 reflected by the depression 20 enters the CCD camera 16.

  When the depression 20 and its periphery 26 in FIG. 20 are photographed by the CCD camera 16 and the intensity of the reflected light received by the CCD camera 16 is measured along a line crossing the depression 20, the result is as shown in FIG.

  Therefore, as shown in FIG. 21, the reflection intensity Ib from the depression 20 is weaker than the reflection light intensity Ia from the periphery 26.

The identification symbol 12 is read by the difference in the intensity of the reflected light intensity.
(C) Misreading occurrence mechanism (i) Form The identification symbol 12 immediately after marking is easy to read. However, as the manufacturing process of the semiconductor device proceeds and various processes are performed on the semiconductor wafer 10, it becomes increasingly difficult to read the identification symbol 12.

  FIG. 22 shows a state in which the manufacturing process has progressed and the wiring 28 is formed in the vicinity of the recess 20 constituting the identification symbol 12. FIG. 23 is a diagram for explaining the result of measuring the reflected light intensity from the surface of the semiconductor wafer 10 along a line crossing the recess 20 and the wiring 28 constituting the identification symbol 12.

  The reflected light intensity Ic from the wiring 28 received by the CCD camera 16 is smaller than the reflected light intensity Ia from the flat wafer surface 32 as shown in FIG. This is because the direction of the reflected light 30 does not become constant when the metal film constituting the wiring 28 is formed in a grain shape or when the thickness of the wiring 28 is approximately the same as the wavelength of light. Conceivable.

  That is, a place (dark place) that appears darkly in places other than the depression 20 appears in the image obtained by photographing the identification symbol 12. As described above, when a dark place appears in a place other than the depression 20 constituting the identification symbol 12, it becomes difficult to interpret the identification symbol 12.

  FIG. 24 shows an example in which an error occurs in reading the identification symbol 12.

  FIG. 24A is an example of an image obtained by observing the surface of the semiconductor wafer immediately after marking the identification symbol 12 with the CCD camera 16. As shown in this figure, the identification symbol 12 immediately after marking is clear and there is no room for error in reading.

  FIG. 24B is an image taken by the CCD camera 16 in a state where the wiring 28 is formed in the region including the identification symbol 12. As shown in the figure, wirings 28 are attached to part of “3” and “C” of the identification symbol 12 consisting of the character string “12345ABC”.

  FIG. 24C shows the result of the image processing unit 18 performing image processing on the image shown in FIG. 24B and reading the identification symbol. As shown in FIG. 24C, the image processing unit 18 misreads the character “3” in the identification symbol “12345ABC” as “8”. Further, the image processing unit 18 misreads the character “C” in the identification symbol “12345ABC” as “0”.

(Ii) Generation Mechanism Next, the reason why such misreading occurs will be described.

  First, a process of reading one character constituting the identification symbol 12 by the image processing unit 18, which is a premise of occurrence of misreading, will be described. FIG. 25 is a diagram for explaining the flow of a process in which the image processing unit 18 reads the characters constituting the identification symbol 12. FIG. 26 is a diagram for explaining image processing performed by the image processing unit 18.

  The image processing unit 18 extracts an image corresponding to a certain character constituting the identification symbol 12 from the image taken by the CCD camera 16 (S10).

  Next, the image processing unit 18 decomposes the image extracted in step S <b> 10 into pixels having a size corresponding to the depression 20. FIG. 26A shows a state in which an area corresponding to the character “3” is extracted from an image obtained by photographing the identification symbol 12 with the CCD camera 16 and the area is decomposed into pixels 34 having a size corresponding to the depression 20. It is a figure explaining.

  Each pixel 34 is a minute region made of a square shown in FIG. The image composed of “●” shown in FIG. 26A is an image before being decomposed into each pixel 34. This image is an image in which the depressions 20 are arranged so as to form the character “3”.

  The pixel 34 covers the entire image. The image processing unit 18 averages the signal intensity (proportional to the reflected light intensity) within each pixel, and binarizes the signal intensity at each pixel based on a predetermined threshold value Ith. That is, the threshold value Ith is set as appropriate, and “0” is associated with a pixel having a signal intensity equal to or greater than the threshold Ith, and “1” is associated with a pixel having a signal intensity equal to or less than the threshold Ith (S20). FIG. 21 shows an example of such a threshold value Ith.

  FIG. 26B shows the result of associating “1” or “0” with each pixel 34 as described above.

  Next, the image processing unit 18 compares the binarized image data as shown in FIG. 26B with the dictionary data recorded in the image processing unit 18 (S30). In the dictionary data, for example, as shown in FIG. 27, data representing each character (for example, “3”) constituting the identification symbol by a combination of pixels having a specific value (for example, “H”) is recorded. ing.

  FIG. 27 shows dictionary data for the character “3”. In FIG. 27, the character “3” is represented by a combination of pixels having the value “H”. The image processing unit 18 sequentially reads out the dictionary data for each character from the dictionary data file provided in itself and compares it with the binarized image data.

  The comparison is performed as follows, for example.

  When the value of the upper left pixel 36 recorded in the dictionary data (FIG. 27) is “H” and the value of the corresponding pixel 38 in the image data (FIG. 26B) is “1”. The values of these pixels 36 and 38 are determined to match. On the other hand, if the value of the image data pixel corresponding to the value “L” of the dictionary data is “0”, it is determined that the two match. Such a determination is performed for all pixels. As a result, when the matching rate (the number of pixels whose values match with respect to the total number of pixels) exceeds a predetermined value (reference value; for example, 90%), the character represented by the dictionary data (in the above example, “ 3 ") is specified as a character for the image data of the comparison partner.

In the example shown in FIGS. 26 and 27, the coincidence rate is 100%, which exceeds the predetermined value (reference value; 90%). Therefore, the character “3” is specified for the extracted image data (FIG. 26A).
Next, the mechanism of misreading will be described by taking as an example image data in which the wiring 28 is applied to the end of the depression 20 constituting the identification symbol “3” as shown in FIG.

  As described above, the wiring 28 appears dark on the image. Therefore, when the pixel 40 corresponding to the wiring 28 is binarized, the value “1” corresponding to the dark place is obtained. Therefore, when the image in which the wiring 28 is applied to the left end of the character “3” shown in FIG. 28A is binarized, the result is as shown in FIG.

  The image data shown in FIG. 28B approximates the character “8” instead of the character “3”.

  FIG. 29 shows dictionary data of the character “3” and the character “8”. FIG. 29A shows dictionary data for the character “3”, and FIG. 29B shows dictionary data for the character “8”.

  The coincidence ratio between the image data shown in FIG. 28B and the dictionary data shown in FIG. 29A (corresponding to the character “3”) is 87% (= 39 ÷ 45 × 100). On the other hand, the matching rate between the image data shown in FIG. 28B and the dictionary data shown in FIG. 29B (corresponding to the character “8”) is 93% (= 42 ÷ 45 × 100).

  Assume that the predetermined value (reference value) for determining the degree of coincidence between the binarized image data and the dictionary data is, for example, 90%. Then, the image processing apparatus 18 specifies the character “8” instead of the correct character “3” with respect to the binarized image data in FIG.

  As described above, when the depression 20 constituting the identification symbol 12 and the surface condition around the depression 20 change due to the progress of the manufacturing process, a darker area than the depression 20 appears in the image photographed by the CCD camera 16. . In such a case, the image processing unit 18 erroneously reads the identification symbol 12. Alternatively, reading the identification symbol itself becomes difficult.

  Note that the cause of misreading of the identification symbol is not limited to the wiring described above, and various structures formed on the semiconductor wafer during the manufacturing process of the semiconductor device cause misreading.

(2) Identification symbol reader provided with oblique illumination unit (Related Art 2)
(A) Configuration FIG. 30 is a schematic view of another identification symbol reader used by the present inventor as viewed from the side. The identification symbol reading device 42 is a device that reads the identification symbol 12 by irradiating the semiconductor wafer 10 with light generated by the oblique illumination unit.

  The identification symbol reading device 42 is used between a plurality of processes for manufacturing a semiconductor device on the semiconductor wafer 10. Then, the identification symbol reader 2 reads the identification symbol 12 made of an array of a plurality of depressions marked on the semiconductor wafer 10 by irradiating this depression and its surroundings with oblique light 44.

  The configuration of the identification symbol reader 42 is as follows.

  First, the identification symbol reading device 42 includes two illuminations 4, and includes a oblique illumination unit 46 that obliquely irradiates oblique light 44 from both sides to and around the plurality of depressions constituting the identification symbol 12.

  In addition, the identification symbol reading device 42 includes a camera 16 that captures an image of an area composed of the depression and its periphery.

Further, the identification symbol reading device 42 includes an image processing unit 18 that processes an image taken by the camera 16 and extracts the identification symbol 12.
(B) Reading Operation Reading the identification symbol 12 using the apparatus shown in FIG. 30 is performed as follows.

  First, the illumination 4 is turned on to irradiate the oblique light 44 to the depressions constituting the identification symbol 12 and the periphery thereof. Here, oblique light refers to light incident on the semiconductor wafer 10 obliquely.

  Next, the identification symbol 12 is photographed by the CCD camera 16, and the image photographed by the CCD camera 16 is processed by the image processing unit 18 to extract the identification symbol 12. That is, the identification symbol 12 is read.

  FIG. 31 is a diagram for explaining a state in which the oblique light 44 is irradiated to the depression 20 constituting the identification symbol 12 and the periphery 26 thereof. Most of the oblique light 44 irradiated to the periphery 26 of the depression 20 is specularly reflected and becomes specularly reflected light 22.

  The regular reflection light 22 is reflected obliquely by the semiconductor wafer 10. By the way, the oblique illumination unit 44 irradiates the semiconductor wafer with oblique light 44 from a high position. Accordingly, the angle θ1 formed by the oblique light 44 and the incoming semiconductor wafer 10 is as large as 60 degrees or more, and therefore the reflection angle θ2 (= θ1) formed by the regular reflected light 22 and the semiconductor wafer 10 is also large. For this reason, most of the regular reflection light 22 is incident on the CCD camera 16 disposed immediately above the semiconductor wafer 10.

  On the other hand, the oblique light 44 irradiated on the depression 20 is reflected (diffusely reflected) in various directions by the curved surface constituting the surface of the depression 20 or the unevenness formed there, and the reflection direction is not constant. For this reason, only a part of the oblique light 44 reflected by the recess 20 enters the CCD camera 16. Accordingly, the depression 20 becomes brighter with respect to the periphery 26 thereof.

  20 is photographed by the CCD camera 16 and the intensity of the reflected light received by the CCD camera 16 with respect to the position coordinate x (see FIG. 31) is measured along a line crossing the depression 20. Then, it becomes like FIG.

  That is, FIG. 21 is a schematic diagram for explaining the intensity of the reflected light incident on the CCD camera along a line traversing the depression 20 and its periphery 26. The horizontal axis is the coordinate x (see FIG. 20) on the line crossing the depression 20, and the vertical axis is the intensity of the reflected light.

  This relationship is the same as the case where the epi-illumination unit 8 described above is used. That is, the reflected light intensity Ia from the periphery 26 of the depression 20 is stronger than the reflection intensity Ib from the depression 20.

  The identification symbol 12 is read by the difference in the intensity of the reflected light intensity.

  As described above, the operation of the identification symbol reading device 42 including the oblique illumination unit 44 is almost the same as that of the epi-illumination unit 8 described above.

Therefore, also in the identification symbol reading device 42 provided with the oblique illumination unit 44, the surface states of the recess 20 and the surrounding 26 constituting the identification symbol 12 are the same as in the identification symbol reading device 2 provided with the above-described illumination unit 8. When changed, an area darker than the surroundings other than the depression 20 appears in the image photographed by the CCD camera 16. For this reason, the image processing unit 18 erroneously reads the identification symbol 12.
(3) Problem Accordingly, an object of the present invention is to provide an identification symbol reading method and an apparatus for reading an identification symbol that is difficult to read due to a change in the state of the surface of the semiconductor wafer in the manufacturing process of the semiconductor device. That is.

(First side)
In order to achieve the above object, according to a first aspect of the present invention, in the identification symbol reading method, a plurality of depressions marked on the semiconductor wafer between a plurality of steps of manufacturing a semiconductor device on the semiconductor wafer. In the identification symbol reading method of reading the identification symbol consisting of the array by irradiating light around the depression and the depression, the traveling direction of the light with respect to the surface of the semiconductor wafer at the position of the identification symbol The method includes a first step of setting an angle formed so that the depression is observed brighter than the surroundings, and a second step of reading the identification symbol after the first step.

  In the first aspect, the angle formed between the semiconductor wafer surface and the traveling direction of the irradiation light is adjusted so that the depression (dot) constituting the identification symbol is observed brighter than the surrounding area. In this way, even if a wiring or the like is formed in the vicinity of the depression constituting the identification symbol, the depression is observed sufficiently brighter than the wiring or the like. That is, even if a structure such as a wiring is formed around the depression (dot), sufficient contrast is maintained between the depression and the surrounding area. Therefore, according to the first aspect, it is possible to provide an identification symbol reading method that prevents misreading of an identification symbol due to the formation of wiring or the like.

(Second side)
To achieve the above object, according to a second aspect of the present invention, in the identification symbol reading method according to the first aspect, the first step includes dictionary data registered in advance as an image of the identification symbol. The angle is adjusted so that the coincidence rate of the image data read as the identification symbol is a predetermined value or more.

  According to the second aspect, it is easy to adjust the angle for making the depression (dot) constituting the identification symbol brighter than the surrounding area.

(Third aspect)
In order to achieve the above object, according to a third aspect of the present invention, in the identification symbol reading method according to the second aspect, the first step includes the light on the semiconductor wafer in addition to the angle. One or both of the intensity and the position where the intensity becomes maximum is adjusted.

  According to the third aspect, the adjustment of the angle for making the dent (dot) constituting the identification symbol brighter than the surroundings is further facilitated.

(Fourth aspect)
In order to achieve the above object, according to a third aspect of the present invention, in the first to third aspects, the first step is a surface state of a region formed by the step and the periphery of the recess. It is performed when the value changes.
(5th side)
In order to achieve the above object, according to a fifth aspect of the present invention, there is provided an identification symbol comprising an array of a plurality of depressions marked on the semiconductor wafer between a plurality of steps of manufacturing a semiconductor device on the semiconductor wafer. In the identification symbol reader for irradiating the recess and the periphery of the recess with light, an illumination unit for irradiating the light, a camera for capturing an area composed of the periphery of the recess and the recess, and the image An image processing unit for processing and extracting the identification symbol, wherein the illumination unit defines an angle formed by the light traveling direction with respect to the semiconductor wafer at the position of the identification symbol from the front periphery. It can be set so that the depression is observed brightly.

  In the fifth aspect, the angle formed between the surface of the semiconductor wafer and the traveling direction of the irradiation light is adjusted so that the depression (dot) constituting the identification symbol is observed brighter than its surroundings. In this way, even if a wiring or the like is formed in the vicinity of the depression constituting the identification symbol, the depression is observed sufficiently brighter than the wiring or the like. That is, even if a structure such as a wiring is formed around the depression (dot), sufficient contrast is maintained between the depression and the surrounding area. Therefore, according to the fourth aspect, it is possible to provide an identification symbol reading device that prevents misreading of identification symbols due to the formation of wiring or the like.

(Sixth aspect)
In order to achieve the above object, according to a sixth aspect of the present invention, in the identification symbol reading device according to the fifth aspect, the image processing apparatus includes a control unit that controls the illumination unit based on the image data. The processing unit creates the image data in which one of the first value corresponding to the bright pixel and the second value corresponding to the dark pixel is associated with each pixel according to the brightness of the pixel. Then, the control unit controls the illumination unit so that the pixel corresponding to the depression has the first value and the pixel corresponding to the surroundings has the second value. Adjust.

  If it approaches to the 6th side, the adjustment of the angle for making the hollow (dot) which comprises an identification symbol brighter than the circumference | surroundings becomes easy.

(Seventh aspect)
To achieve the above object, according to a seventh aspect of the present invention, an identification symbol consisting of an array of a plurality of depressions marked on a semiconductor wafer is read by irradiating light around the depressions and the depressions. A first step and a second step of setting an angle formed by the light traveling direction with respect to the surface of the semiconductor wafer at the position of the identification symbol so that the depression is observed brighter than the surroundings. And a third step of reading the identification symbol after the second step.

  In the present invention, the angle formed between the surface of the semiconductor wafer and the traveling direction of the irradiation light is adjusted so that the depression (dot) constituting the identification symbol is observed brighter than the surrounding area. In this way, even if a wiring or the like is formed in the vicinity of the depression constituting the identification symbol, the depression is observed sufficiently brighter than the wiring or the like. That is, even if a structure such as a wiring is formed around the depression (dot), sufficient contrast is maintained between the depression and the surrounding area. Therefore, according to the present invention, it is possible to provide an identification symbol reading method and apparatus for preventing an identification symbol from being misread due to the formation of a wiring or the like.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the technical scope of the present invention is not limited to these embodiments, but extends to the matters described in the claims and equivalents thereof. Note that even if the drawings are different, corresponding parts are denoted by the same reference numerals, and description thereof is omitted.

(Embodiment 1)
In the present embodiment, the identification symbol 12 can be read even when the surface state of the region composed of the depression 20 and the periphery 26 constituting the identification symbol 12 changes with the progress of the manufacturing process of the semiconductor device. The present invention relates to an identification symbol reader and a reading method.

(1) Identification Symbol Reading Device The identification symbol reading device in the present embodiment has an illumination unit that can be set to a small angle (for example, smaller than 45 degrees) between the semiconductor wafer surface and the traveling direction of irradiation light. . For this reason, it is possible to reduce the reflected light that enters the camera from a region other than the depression (dot) constituting the identification symbol. For this reason, even if a wiring or the like is formed in the vicinity of the identification symbol, the depression and its surrounding contrast can be kept sufficiently wide, so that misreading due to the formation of the wiring or the like can be prevented.

(I) Configuration FIG. 1 is a schematic view of an identification symbol reading device 54 in this embodiment as viewed from the side. FIG. 2 is a schematic diagram showing a bird's-eye view of the main part of the identification symbol reading device 54 in the present embodiment.

  The identification symbol reader 54 is used between a plurality of processes for manufacturing a semiconductor device (not shown) on the semiconductor wafer 10. Then, the identification symbol reading device 54 reads the identification symbol 12 marked on the semiconductor wafer 10 and made up of an array of a plurality of depressions (dots) by irradiating the depression 58 and its surroundings with light 58. In addition, the said hollow (dot) is formed by laser marking, for example.

  The configuration of the identification symbol reader 54 is as follows.

  First, the identification symbol reading device 54 includes an illumination unit 4 made of, for example, a light emitting diode (LED), and includes an illumination unit 48 that irradiates light 58 to a plurality of depressions constituting the identification symbol 12 and the periphery thereof.

  Further, the identification symbol reading device 54 includes a camera 50 (for example, a CCD camera) that captures an image of an area including the depression and the periphery thereof.

  Further, the identification symbol reading device 54 includes an image processing unit 18 that processes an image captured by the camera 50 and extracts the identification symbol 12.

  Further, as shown in FIG. 1, the identification symbol reading device 54 may include a control unit 52 that controls the illumination unit 48 based on image data captured by the camera 50.

The control unit 52 is not always necessary. When the control unit 52 is not used, the operator may adjust the illumination unit 48 based on the image data.
(II) Reading Operation (i) Operation The identification symbol 12 is read as follows by the apparatus shown in FIG.

  First, the illumination unit 48 is turned on, and the light 58 is irradiated on the depressions constituting the identification symbol 12 and the surroundings thereof.

  Here, the angle θ formed by the traveling direction of the light 58 at the position of the identification symbol 12 (the direction in which the light intensity becomes maximum) with respect to the surface of the semiconductor wafer 10 is defined as a depression (dot) constituting the identification symbol 12. 20 is set to be observed brighter than its surroundings 26. For this purpose, for example, the light 58 is incident on the surface of the semiconductor wafer 10 at an angle θ of less than 45 degrees and greater than 0 degrees.

  As will be described later, the smaller the angle θ formed between the irradiation light 58 and the semiconductor wafer 10, the less the reflected light from the periphery of the recess constituting the identification symbol, and the brighter the recess. Even if a wiring or the like is formed in the vicinity of the identification symbol in such a state, the image of the wiring or the like is sufficiently darker than the depression. For this reason, even if a wiring or the like is formed around the depression, the contrast between the depression and the surrounding area is kept sufficiently wide. Accordingly, misreading of the identification symbol is less likely to occur.

  After such adjustment of the irradiation light angle, the identification symbol 12 is photographed by the camera 50, and the image photographed by the camera 50 is processed by the image processing unit 18 to extract the identification symbol 12. That is, the identification symbol 12 is read.

  By the way, in the identification symbol reader 2 using the incident light described above, the angle θ formed between the irradiation light and the semiconductor wafer is 90 degrees. Also in the identification symbol reader 42 using oblique light, the angle θ is 60 degrees or more. That is, the angle θ of the light 58 used in the identification symbol reading device 54 in this embodiment is lower than those devices. Therefore, hereinafter, the light irradiated onto the semiconductor wafer 10 at such a low angle will be referred to as low oblique light.

(Ii) Principle First, an image photographed by the camera 50 when a wiring or the like is formed in the depression 20 constituting the identification symbol 12 and its periphery 26 will be described.

  FIG. 3 is a diagram for explaining a state in which the low oblique light 58 is irradiated to the depression 20 constituting the identification symbol 12 and its periphery 26 by the identification symbol reading device 54 according to the present embodiment. Here, FIG. 3 shows a state where wirings and the like are not yet formed in the recess 20 constituting the identification symbol 12 and its periphery 26.

  The low oblique light 60 irradiated to the periphery 26 of the depression 20 is specularly reflected on the flat semiconductor wafer surface to become specularly reflected light 22. Here, the angle θ between the semiconductor wafer 10 and the low oblique light 60 is a low angle smaller than 45 degrees, for example. Accordingly, the angle φ (= θ) formed by the regular reflected light 22 with the surface of the semiconductor wafer 10 is also reduced (becomes small), so that almost no light is incident on the camera 50.

  On the other hand, the low oblique light 62 irradiated to the depression 20 is reflected (diffusely reflected) in various directions by the curved surface constituting the surface of the depression 20 or the unevenness formed there. For this reason, a part of the low oblique light 64 reflected by the recess 20 enters the camera 50. For this reason, the depression 20 becomes brighter than the surrounding area 26.

  FIG. 4 is a schematic diagram illustrating the intensity of the reflected light incident on the camera 50 along a line traversing the depression 20 and its periphery. The horizontal axis is the coordinate x (see FIG. 3) on the line crossing the depression 20, and the vertical axis is the intensity of the reflected light.

  Contrary to the related art in which the identification symbol is irradiated with incident light or oblique light, the reflected light intensity Ib ′ generated in the recess 20 is stronger than the intensity Ia ′ of the reflected light generated around the periphery 26 of the recess 20. The identification symbol 12 is read by the difference in the intensity of the reflected light intensity.

  Next, an image photographed by the camera 50 when a wiring or the like is formed in the depression 20 constituting the identification symbol 12 and its periphery 26 will be described.

  FIG. 5 is a diagram for explaining a state in which the low oblique light 60, 62 is irradiated to the depression 20 constituting the identification symbol 12 and its surroundings 26 by the identification symbol reading device 54 in the present embodiment. FIG. 5 shows a case where wiring is formed in a region including the depression 20 constituting the identification symbol 12 and its periphery 26.

  The low oblique light 60 irradiated to the periphery 26 of the depression 20 is specularly reflected on the flat semiconductor wafer surface to become specularly reflected light 22. Here, the angle θ between the semiconductor wafer 10 and the low oblique light 60 is a low angle smaller than 45 degrees, for example. Accordingly, the angle φ (= θ) formed by the regular reflection light 22 with the surface of the semiconductor wafer 10 is also low, and most of the light does not enter the camera 50.

  On the other hand, the low oblique light 62 irradiated to the depression 20 is reflected (diffuse reflection) in various directions by the curved surface constituting the surface of the depression 20 or the unevenness formed there, and the reflection direction is not constant. For this reason, a part of the low oblique light 64 reflected by the recess 20 enters the camera 50. Accordingly, the depression 20 becomes brighter than the surrounding area 26.

  Further, when the metal film constituting the wiring 28 is formed in a grain shape, or when the thickness of the wiring 28 is approximately the same as the wavelength of light, the direction of the reflected light 30 reflected by the wiring 28 is also constant. No longer. For this reason, a part of the reflected light 30 enters the camera 50, and the wiring 28 becomes brighter than the periphery 26 of the depression. However, since the surface of the wiring 28 is flatter than the inside of the depression 20, the wiring 28 is darker than the depression 20. That is, the order of brightness is the order of the surface of the semiconductor wafer 10 in which the depression 20, the wiring 28, the depression 20 and the wiring 28 are not formed (the depression 20 is the brightest).

  FIG. 6 is a schematic diagram illustrating the intensity of the reflected light incident on the camera 50 along a line traversing the depression 20 and its periphery. The horizontal axis is the coordinate x (see FIG. 3) on the line crossing the depression 20, and the vertical axis is the intensity of the reflected light.

  The reflected light intensity Ib ′ generated in the recess 20 is the strongest, the reflected light intensity Ic ′ generated next in the wiring 28 is the strongest, and the reflected light from the region 32 around the recess 20 where the wiring 28 is not formed. Is the weakest.

  As described above, since the angle θ between the low oblique light 60 irradiated to the semiconductor wafer 10 and the semiconductor wafer 10 is small, the intensity Ia ′ of the reflected light incident on the camera 50 from the periphery 26 of the recess 20 constituting the identification symbol 12. Is extremely small.

  If the wiring 28 or the like is formed around the depression 26, the direction of the reflected light is disturbed, and a part of the reflected light enters the camera 50. For this reason, the intensity Ic ′ of the reflected light that enters the camera 50 from the wiring 28 is stronger than the surroundings of the wiring 28 and the like. However, since the surfaces of the wirings 28 and the like are kept flat to some extent, the wirings 28 are reflected toward the camera 50 when the angle θ is reduced as in the periphery 32 of the wirings 28 and the like (a flat region on the surface of the semiconductor wafer). The amount of light decreases rapidly.

  On the other hand, the intensity of the reflected light entering the camera 50 from the depression 20 does not depend much on the angle θ.

  That is, when the angle θ formed by the low oblique light 60 and the semiconductor wafer 10 is reduced, the image of the depression 20 becomes brighter than the flat surface 32 of the semiconductor wafer 10 (around the depression where the wiring 28 is not formed).

  Therefore, even if the wiring 28 and the like are formed around the recess 20, the recess 20 whose inside is greatly curved and fine irregularities are formed on the surface reflects much more light to the camera 50 than the surroundings. Therefore, the image processing unit 18 can easily read the identification symbol 12.

  Therefore, according to the identification symbol reading device of the present embodiment, even identification symbols that have been difficult to read conventionally due to a change in the state of the semiconductor wafer surface accompanying the manufacture of the semiconductor device can be easily read. be able to. That is, no misreading occurs even if a wiring or the like is formed in the vicinity of the identification symbol 12.

(2) Identification Symbol Reading Method As is apparent from the above description, illumination is performed even when the surface state of a semiconductor wafer changes with the manufacture of a semiconductor device, making it difficult to read the identification symbol. By using low oblique light, the identification symbol 12 marked on the semiconductor wafer can be read.

  FIG. 7 is a diagram illustrating the steps of a method for reading the identification symbol by irradiating the identification symbol 12 with low oblique light.

  The method for reading the identification symbol described here is that the identification symbol consisting of an array of a plurality of depressions marked on the semiconductor wafer is divided between the plurality of steps for manufacturing the semiconductor device on the semiconductor wafer. This is a method of reading by irradiating with light.

  First, the angle formed by the light traveling direction with respect to the surface of the semiconductor wafer at the position of the identification symbol is set so that the depression is observed brighter than its surroundings (S10).

  Here, the angle is preferably smaller than 45 degrees and larger than 0 degrees. More preferably, it is smaller than 30 degrees and larger than 5 degrees, and most preferably smaller than 25 degrees and larger than 10 degrees. This is because, as is clear from the above description, the smaller the angle θ between the low oblique light 58, 60, 62 and the semiconductor wafer 10, the greater the contrast between the depression 20 and its surroundings. However, if the angle θ is too small, it is difficult for light to reach the recess 20, and it is not preferable that the angle θ is too small.

  Thereafter, the identification symbol is read (S20).

  In this way, even an identification symbol that has conventionally been difficult to read due to a change in the state of the surface of the semiconductor wafer accompanying the manufacture of the semiconductor device can be read. That is, no misreading occurs even if a wiring or the like is formed in the vicinity of the identification symbol 12.

(Embodiment 2)
This embodiment is an example of an identification symbol reading apparatus in which the image processing unit 18 and the control unit 52 of FIG.
(1) Configuration The configuration 54 of the identification symbol reader in this embodiment is basically the same as the identification symbol reader 54 described in the first embodiment (see FIGS. 1 and 2).

  However, in the present embodiment, the image processing unit 18 has one of a first value (for example, “1”) corresponding to a bright pixel and a second value (for example, “0”) corresponding to a dark pixel. Image data corresponding to each pixel is created according to the brightness of each pixel.

  In the present embodiment, the control unit 52 controls the illumination unit 48 so that the pixel corresponding to the depression 20 has the first value (for example, “1”) and corresponds to the periphery 26 of the depression 20. The angle θ formed by the traveling direction of the low oblique light 58 with respect to the semiconductor wafer 10 at the position where the identification symbol 12 is marked is adjusted so that the pixel to be processed has a second value (for example, “0”).

(2) Operation (i) Operation of Image Processing Unit Image processing is performed in the case where the angle θ formed by the low oblique light 58 with respect to the semiconductor wafer 10 is already set so that the depression 10 becomes brighter than its surroundings. The operation of the unit 18 will be described. However, even if the angle θ is not set in this way, there is no difference in the operation itself performed by the image processing unit 18.

  The image processing unit 18 reads the entire identification symbol 12 by reading characters constituting the identification symbol 12 one by one.

  The process of reading the characters constituting the identification symbol 12 by the image processing unit 18 is basically the same as the process of reading the characters by the image processing unit in the related art shown in FIG. FIG. 8 is a diagram illustrating image processing performed by the image processing unit 18.

  First, the image processing unit 18 extracts an image corresponding to a certain character constituting the identification symbol 12 from the image taken by the camera 50 (S10).

  Next, the image processing unit 18 divides the image extracted in step S <b> 10 into pixels having a size corresponding to the depression 20. FIG. 8A illustrates a state in which an area corresponding to the character “3” is extracted from an image obtained by capturing the identification symbol 12 with the camera 50 and the area is decomposed into pixels 34 having a size corresponding to the depression 20. It is a figure to do.

  Each pixel 34 is a minute area formed of a square shown in FIG. The image composed of “◯” shown in FIG. 8A is an image of the depression 20 before being decomposed into each pixel 34. Here, the depression 20 is bright and the surrounding area is dark.

  The pixel 34 covers the entire image. The image processing unit 18 averages the signal intensity (proportional to the reflected light intensity) within each pixel, and binarizes the signal intensity at each pixel based on a predetermined threshold value I′th. That is, a threshold value I′th is appropriately set, and “1” is associated with a pixel having a signal intensity equal to or higher than the threshold I′th, and “0” is assigned to a pixel with a signal intensity equal to or lower than the threshold I′th. Correspond (S20). FIG. 4 shows an example of such a threshold value I′th.

  FIG. 8B shows the result of associating “1” or “0” with each pixel 34 as described above.

  Next, the image processing unit 18 compares the binarized image data as shown in FIG. 8B with the dictionary data recorded in the image processing unit 18 and derives a matching rate (S30). In the dictionary data, for example, as shown in FIG. 27, each character (for example, “3”) constituting the identification symbol is represented by a combination of pixels corresponding to the value “H” (or “L”). Is recorded.

  FIG. 27 shows dictionary data for the character “3”. In FIG. 27, the character “3” is represented by a combination of pixels corresponding to the value “H”. The image processing unit 18 sequentially reads out the dictionary data for each character from the dictionary data file provided for itself and compares it with the binarized image data.

  For example, when the value of the upper left pixel 36 in the dictionary data (FIG. 27) is “H” and the value of the corresponding pixel 38 in the image data (FIG. 8B) is “1”. Determines that the values of these pixels 36 and 38 match. Such a determination is performed for all pixels. As a result, when the matching rate (the number of pixels having a matching value with respect to the total number of pixels) exceeds a predetermined value (for example, 90%), the character represented by the dictionary data (in the above example, “3”) is changed. It is specified as a character for the image data being compared (FIG. 8B).

  In the example shown in FIGS. 8 and 27, since the matching rate is 100% and exceeds the predetermined value (90%), the character “3” is specified for the extracted image data.

  The table shown in FIG. 9 is a table in which the matching rate between the image data in FIG. 8 and the dictionary data in FIG. 27 is compared for each row having the same row number. For example, as shown in this table, the image processing unit 18 calculates a matching rate (horizontal matching rate) for each row, and determines the overall matching rate based on the result.

  The example shown in FIG. 8 is an example in which no wiring or the like is formed around the depression 20 constituting the identification symbol 12 and its periphery. Next, the reading operation with respect to the image in which the recesses 20 constituting the identification symbol 12 and wirings around them are formed will be described. In the example shown in FIG. 10, the wiring 28 is applied to the end of the array (set) of the depressions 20 constituting the identification symbol “3”.

  As described above, the image of the periphery 26 of the depression taken by the camera 50 is darkly displayed including the area where the wiring 28 is formed (see FIG. 10).

  Similarly, as described above, the region where the wiring 28 is formed is slightly brighter than the other surroundings 26. However, the image of the wiring 28 is sufficiently darker than the depression 20. Therefore, as shown in FIG. 6, the intensity I′a, I′c of the image signal with respect to the periphery of the depression 20 is smaller than the appropriately set threshold value I′th. The depression 20 is photographed sufficiently brightly even if the wiring 28 is applied.

  Therefore, the image processing unit 18 associates the pixel where the depression 20 is formed with the numerical value “1” corresponding to the bright place and the pixel corresponding to the periphery 26 of the depression corresponds with the numerical value “0” corresponding to the dark place. Can be made.

  FIG. 10B shows the result of associating each pixel 34 with “1” or “0” as described above. This figure is the same as FIG. 8B, which shows the result when no wiring or the like is formed in the depression 20 and its periphery 26.

  Accordingly, the image processing unit 18 can extract the character “3” without error as in the case where no wiring or the like is formed in the depression 20 and its surroundings 26.

(Ii) Operation of Control Unit The control unit 52 of the present embodiment controls the illumination unit 48 so that, for example, in the identification symbol 12, the angle θ formed by the traveling direction of the low oblique light 58 with respect to the semiconductor wafer 10 is For example, adjustment is made so as to obtain an optimum value derived in advance. Here, the optimum value derived in advance is that the pixel corresponding to the depression 20 is assigned the first value (“1” for the bright place), and the pixels corresponding to the periphery 26 of the depression 20 are the second values. It is an angle θ (for example, 40 degrees) that makes it possible to assign a value (“0” with respect to a dark place).

  The lighting unit 48 in the present embodiment includes, for example, a light emitting member 66 that includes a plurality of LEDs 4 and the intensity of emitted light is biased in one direction (that is, a light emitting member that emits light having high directivity), and this The rotating member 68 is configured to rotate the light emitting member 66 so that the direction in which the intensity of the emitted light (low oblique light) from the light emitting member 66 is the strongest (direction indicating the directivity) changes the angle θ formed with the semiconductor wafer 10. Yes.

  The control unit 52 controls the rotation angle of the rotating member 68 to adjust the angle θ formed by the low oblique light 58 and the semiconductor wafer 10.

  Therefore, according to the identification symbol reading device in the present embodiment, even identification symbols that have been difficult to read conventionally due to a change in the state of the semiconductor wafer surface accompanying the manufacture of the semiconductor device can be easily read. become able to. That is, no misreading occurs even if a wiring or the like is formed in the vicinity of the identification symbol 12.

(Embodiment 3)
In the present embodiment, the identification symbol 12 can be accurately read even when the surface state of the region composed of the recess 20 and the periphery 26 constituting the identification symbol 12 changes with the progress of the manufacturing process of the semiconductor device. The present invention relates to an identification symbol reading apparatus and an identification symbol reading method provided with a control unit 52 that automatically adjusts the illumination unit in accordance with the change in the surface state.
(1) Identification Symbol Reading Device (I) Configuration FIG. 11 is a schematic view of the identification symbol reading device 54 in the present embodiment as viewed from the side.

  The configuration of the identification symbol reader in the present embodiment is basically the same as that of the identification symbol reader in the first embodiment shown in FIGS.

  However, in the present embodiment, in the identification symbol reader 54 of the first or second embodiment, the control unit 52 controls the illumination unit 50 so that the dictionary data registered in advance as an image of the identification symbol 12 The angle formed by the traveling direction of the low oblique light 58 with respect to the semiconductor wafer 10 at the position of the identification symbol 12 so that the coincidence rate of the image data read as the identification symbol is a predetermined value (for example, 90%) or more. Adjust.

  Further, in the identification symbol reading device 54, the control unit 52 controls the illumination unit 48 to adjust the intensity of low oblique light at the position where the identification symbol is marked, in addition to the angle.

  FIG. 12 shows an example of the illumination unit 48 provided in the identification symbol reading device 54 in the present embodiment. The illumination unit 48 of the present embodiment is basically the same as the illumination unit 48 of the first and second embodiments. However, the configuration of the illumination unit 48 of the present embodiment is more specific.

  FIG. 12 shows two illumination units 48 that irradiate the semiconductor wafer 12 with low oblique light from the left and right. 12A is a view of the identification symbol reading device 48 as seen from the side, and FIG. 12B is a view as seen from the front.

  The illumination unit 48 illustrated in FIG. 12 includes, for example, a light emitting member 66 including a plurality of LEDs 4 arranged in a grid pattern and a plate-like support body on which the LEDs 4 are arranged. Here, the light emitted from the LEDs 4 has directivity, and each LED 4 is mounted so that the direction 70 perpendicular to the support coincides with the direction in which the intensity of the emitted light reaches a maximum value.

  Here, each LED 4 belongs to one of a plurality of rows in which the LEDs 4 are aligned in a direction parallel to the surface of the semiconductor wafer 10. In the example shown in FIG. 12B, each LED 4 belongs to any one of the a column, the b column, and the c column.

  Further, the illumination unit 48 includes a rotating member 68 that rotates the light emitting member 66 about an axis 72 parallel to the semiconductor wafer 10.

  Therefore, according to the illumination unit 48 of the present embodiment, the angle θ formed by the traveling direction of the low oblique light emitted from the illumination unit 48 with respect to the semiconductor wafer 10 is adjusted by adjusting the rotation angle of the rotating member 68. It is possible.

  Further, the lighting unit 48 controls the lighting of the LEDs 4 for each row in which the LEDs 4 are arranged, whereby the intensity of low oblique light at the position where the identification symbol is marked can be adjusted.

(II) Operation Next, the operation of the identification symbol reading device 54 in the present embodiment will be described.

  The operation of the identification symbol reader 54 in this embodiment is basically the same as the operation of the identification symbol reader of the first embodiment.

  That is, the identification symbol reader 54 first determines the angle θ formed by the traveling direction of the low oblique light 58 (the direction in which the light intensity is maximum) at the position of the identification symbol 12 with respect to the surface of the semiconductor wafer 10. 12 is set so that the depression 20 constituting 12 is observed brighter than the surroundings 26, 28, 32 (first step).

Thereafter, the identification symbol is read (second step).
However, in the first step, the identification symbol reading device 54 in the present embodiment is the dictionary data registered in advance as the image of the identification symbol 12 by the control unit 52 and the image read as the identification symbol. The angle θ is adjusted so that the data coincidence rate becomes a predetermined value (for example, 90% or more).

  Further, the identification symbol reader 54 adjusts the intensity of low oblique light at the position where the identification symbol is marked, in addition to the angle θ.

  As shown in FIG. 11, the identification symbol reader 54 can set four types of angles A, B, C, D, and E as the rotation angles of the illumination unit 48. Further, the identification symbol reading device 54 can control lighting for each row of LEDs 4 (a row, b row, and c row; hereinafter referred to as a lighting row) constituting the light emitting member 66 of the illumination unit 48 (see FIG. 12). ).

  The identification symbol reader 54 adjusts the angle θ formed by the low oblique light 58 and the semiconductor wafer 10 according to the rotation angle of the illumination unit 48 (A, B, C, D, E; hereinafter referred to as oblique illumination angle). Further, the identification symbol reading device 54 adjusts the intensity of low oblique light at the position where the identification symbol is marked by the columns of the LEDs 4 of the illumination unit 48 (a column, b column, and c column). That is, the control unit 52 sets illumination conditions including combinations of various oblique illumination angles and lighting rows for the illumination unit 48, and controls the illumination unit 48 based on the illumination conditions.

  FIG. 13 shows the results of measuring the coincidence rate between image data (for example, FIG. 8B) and dictionary data (FIG. 27) by variously changing the combination of oblique illumination angle and lighting row (irradiation conditions). It is a table. In the first column of the table shown in FIG. 13, the oblique illumination angle set by the control unit 52 is described. In the second column of the table shown in FIG. 13, lighting positions set by the control unit 52 are described. This table is recorded in a memory (not shown) constituting the image processing unit 18.

  The coincidence rate obtained for the combination of the oblique illumination angle and the lighting column in a certain row is described in the third column of the same row.

  FIG. 14 shows an angle θ formed between the semiconductor wafer 10 and the low oblique light 58 and the low oblique light 58 so that the coincidence rate between the image read as the characters constituting the identification symbol and the dictionary data is equal to or higher than a predetermined value (reference value). It is a figure explaining a specific example of the 1st step which adjusts the intensity of.

  First, the control unit 52 sets the first irradiation condition (for example, the combination in the third row in FIG. 13; the oblique illumination angle is A and the lighting columns are a, b, and c) (S10).

  Next, based on this irradiation condition, the illumination unit 48 is controlled, and the light emitting member 66 is turned on (S20).

  Next, the image processing unit 18 processes an image of a portion of the identification symbol 12 illuminated by the low oblique light 58 from the illumination unit 48 and knows what the marked character is, and dictionary data Is measured and recorded (S30).

  Next, the control unit 52 sets the next irradiation condition (for example, the combination in the fourth row in FIG. 13; the oblique illumination angle is A and the lighting columns are a and b) in the memory (S40). However, when all the irradiation conditions are set and there is no new irradiation condition, no new irradiation condition is set.

  Next, the control unit 52 determines whether the measurement of the coincidence rate has been completed for all irradiation conditions (the measurement of the coincidence rate is scheduled) (S50).

  If the measurement of all coincidence rates has not been completed, the process returns to S20, and the coincidence rate under the next irradiation condition is measured. Such a procedure is repeated and the coincidence rate under all irradiation conditions is measured.

  FIG. 14 is a table that records combinations of coincidence rates (third column) and measurement conditions (first column and second column) measured by such processes.

  When the measurement of the coincidence rate under all irradiation conditions is completed, the control unit 54 selects a measurement condition that exceeds a predetermined value (reference value; for example, 90%) and achieves the highest coincidence rate. And according to the selected conditions, the rotation angle and lighting row | line | column of the illumination unit 48 are set, and the light emission member 66 is lighted (S60).

  Thereafter, the process proceeds to a second step of reading the entire identification symbol marked on the semiconductor wafer 10. FIG. 15 is a diagram for explaining a specific example of the second step of reading the entire identification symbol marked on the semiconductor wafer 10.

  First, the image processing unit 18 reads the identification symbol (ID) marked on the semiconductor wafer 10 (S70).

  Next, among the characters constituting the read identification symbol (ID), the matching rate is compared with a predetermined value (reference value; for example, 90%) for characters whose markings are known in advance (S80). ).

If the coincidence rate is equal to or higher than the predetermined value, it is determined whether reading of the identification symbols has been completed for all the semiconductor wafers (S90).
If reading of all the semiconductor wafers has been completed, the reading operation is finished.

  If reading for all the semiconductor wafers has not been completed, the next semiconductor wafer is set (S100). Then, it returns to step S70 and the identification symbol of the other semiconductor wafer which has not yet read the identification symbol is read.

  On the other hand, when the coincidence rate is less than a predetermined value (reference value), the process returns to step S10, and the first step is performed again on the semiconductor wafer whose coincidence rate is equal to or less than the predetermined value to search for preferable irradiation conditions. Through such a process, an irradiation condition in which the coincidence rate is equal to or higher than a predetermined value (reference value) is found.

  The identification symbol is usually read for each carrier in which a plurality of semiconductor wafers that have undergone similar manufacturing processes are stored. For this reason, for semiconductor wafers stored in the same carrier, the irradiation conditions for the coincidence rate to be equal to or higher than a predetermined value are considered to be substantially the same.

  Therefore, it is beneficial in terms of work efficiency to continue reading the identification symbol under the previously set irradiation condition until the matching rate does not satisfy the predetermined value (reference value) as in the second step (FIG. 15). is there.

  In other words, when the surface state of the region composed of the recess 20 and its periphery 26 is changed by the manufacturing process of the semiconductor device, the identification symbol can be read efficiently by performing the first step again. .

  Next, the step of deriving the matching rate (S30) will be described in detail based on a specific example.

  FIG. 16 is a diagram for explaining the flow of the step (S30) in which the image processing unit 18 derives the matching rate. 17 and 18 are diagrams for explaining image processing performed by the image processing unit 18.

  The image processing unit 18 extracts an image for one character from an area that is known in advance from the image photographed by the camera 50 (S10).

  Next, the image processing unit 18 decomposes the image extracted in step S10 into pixels 34 having a size corresponding to the depressions 74 and 76. 17A, an area corresponding to the known character “3” is extracted from an image obtained by photographing the identification symbol with the camera 50, and the area is a pixel 34 having a size corresponding to the depressions (dots) 74 and 76. Shows the disassembled state.

  The image shown in FIG. 17A is an image in a state where the wiring 28 is applied to the end of the array of the depressions 20 representing the character “3”. The irradiation condition of the low oblique light 58 is the condition shown in the fifth line of FIG. That is, the oblique illumination angle is A, and the lighting row is c.

  In this case, since the angle θ formed by the low oblique light 58 and the semiconductor wafer 10 is sufficiently low, the periphery 26 of the depression (dot) and the wiring 28 are darker than the threshold value I′th (see FIGS. 5 and 6). On the other hand, a part of the depressions 74 irradiated with the low oblique light 58 becomes brighter than the threshold value I′th because the low oblique light 58 is irregularly reflected inside the depressions 74.

  However, since the curvature and state of the depression surface are not uniform, a dark depression 76 appears if the amount of low oblique light 58 is not sufficient. FIG. 17A shows such a state.

  The pixel 34 covers the entire image. The image processing unit 18 averages the signal intensity (proportional to the reflected light intensity) within each pixel, and binarizes the signal intensity at each pixel based on a predetermined threshold value (reference value) I′th. Here, “1” is made to correspond to a pixel whose average signal intensity is equal to or higher than the threshold I′th, and “0” is made to correspond to a pixel whose average signal intensity is equal to or lower than the threshold I′th. (S20).

FIG. 17B shows the result of associating “1” or “0” with each pixel 34 as described above.
Next, the image processing unit 18 compares the binarized image data as shown in FIG. 17B with the dictionary data recorded in the image processing unit 18 (S30). In the dictionary data, for example, as shown in FIG. 27, data representing each character (for example, “3”) constituting the identification symbol by a combination of pixels having a specific value (for example, “H”) is recorded. ing.

  FIG. 27 shows dictionary data for the character “3”. In FIG. 27, the character “3” is represented by a combination of pixels having the value “H”. The image processing unit 18 sequentially reads out the dictionary data for each character from the dictionary data file provided in itself and compares it with the binarized image data.

  For example, when the value of the upper left pixel 36 in the dictionary data (FIG. 27) is “H” and the value of the corresponding pixel 38 in the image data (FIG. 26B) is “1”. Determines that the values of these pixels 36 and 38 match. Such a determination is performed for all the pixels, and the coincidence rate (the number of pixels whose values coincide with the total number of pixels) is derived.

  The coincidence rate can be obtained, for example, by obtaining the coincidence rate (lateral coincidence rate) of pixels belonging to the same row for each row and averaging the lateral coincidence rates (see FIGS. 17 and 18).

Therefore, according to the identification symbol reading device in the present embodiment, even identification symbols that have conventionally been difficult to read due to a change in the state of the semiconductor wafer surface accompanying the manufacture of the semiconductor device can be read. become. That is, even if a wiring or the like is formed in the vicinity of the identification symbol 12, misreading does not occur.
(2) Identification Symbol Reading Method The identification symbol reading method in the present embodiment is basically the same as the identification symbol reading method described in the first embodiment (see FIG. 7).

  However, in the present embodiment, in the first step (S10), dictionary data (dictionary data) registered in advance as an image of an identification symbol and image data read as an identification symbol (binarized image data) The angle θ formed by the semiconductor wafer 10 and the low oblique light 58 is adjusted such that the coincidence ratio of the semiconductor wafer 10 and the low oblique light 58 is equal to or greater than a predetermined value (reference value; for example, 90%).

  Furthermore, in the first step (S10), in addition to the angle θ, the intensity of low oblique light (at the position where the identification symbol is marked) may be adjusted.

  Adjustment of the angle θ is performed by the control unit 18 controlling the rotating unit 68 of the illumination unit 48. The intensity of the low oblique light (at the position where the identification symbol is marked) is determined by the control unit 18 controlling the lighting row of the light emitting members 66 of the illumination unit 48.

  Therefore, according to the identification symbol reading device in the present embodiment, even identification symbols that have been difficult to read conventionally due to a change in the state of the surface of the semiconductor wafer accompanying the manufacture of the semiconductor device can be read very easily. Can do.

  Such a method for reading the identification symbol is effective when the surface state of the region composed of the depression and the periphery of the depression changes due to the manufacturing process of the semiconductor device.

  The identification symbol is not limited to a combination of characters, and may be a barcode, for example. Further, the identification symbol may be formed not by laser marking but by, for example, etching a semiconductor wafer using a photoresist film.

  The present invention can be used in the semiconductor device manufacturing industry.

It is the schematic which looked at the identification symbol reading apparatus in Embodiment 1 from the side. FIG. 3 is a schematic diagram bird's-eye view of a main part of the identification symbol reading device according to the first embodiment. It is a figure explaining the state which irradiated the low oblique light to the hollow which comprises an identification symbol, and its circumference | surroundings (when wiring etc. are not formed in a hollow and its circumference | surroundings). In Embodiment 1, it is the schematic explaining the intensity | strength of the reflected light which injects into a camera along the line which traverses the hollow which comprises the identification symbol, and its circumference in Embodiment 1 (a wiring etc. are formed in a hollow and its circumference) If not). It is a figure explaining the state which irradiated the low oblique light to the hollow which comprises an identification symbol, and its circumference | surroundings (when wiring etc. are formed in the hollow and its circumference | surroundings). In Embodiment 1, it is the schematic explaining the intensity | strength of the reflected light which injects into a camera along the line which traverses the hollow which comprises the identification symbol, and its circumference in Embodiment 1 (a wiring etc. are formed in a hollow and its circumference) If so). FIG. 5 is a diagram illustrating a process of an identification symbol reading method in the first embodiment. FIG. 10 is a diagram for explaining image processing executed by the image processing unit according to the second embodiment (part 1); It is the table | surface which compared the coincidence rate of image data and dictionary data for every line. FIG. 10 is a diagram for explaining image processing executed by the image processing unit according to the second embodiment (part 2); It is the schematic which looked at the identification symbol reading apparatus in Embodiment 3 from the side. It is a figure explaining the structure of the illumination unit in Embodiment 3. FIG. It is a figure explaining the result of having changed the combination (irradiation conditions) of the oblique illumination angle in an illumination unit variously, and measuring the coincidence rate of image data and dictionary data. First step of the identification symbol reading operation in the third embodiment (the angle θ of the low oblique light so that the coincidence rate between the image read as the characters constituting the identification symbol and the dictionary data is equal to or higher than a predetermined value. And a step of adjusting the irradiation light intensity). It is a figure explaining a specific example of the 2nd step (step which reads the whole identification symbol marked on the semiconductor wafer) of the identification symbol reading operation in Embodiment 3. FIG. It is a figure explaining the flow from which an image processing unit derives a coincidence rate. It is FIG. (1) explaining the image processing performed by an image processing unit. It is FIG. (2) explaining the image processing performed by an image processing unit. It is the schematic which looked at the identification symbol reading apparatus provided with the epi-illumination unit from the side. It is a figure explaining the state which irradiated the incident light on the hollow which comprises an identification symbol, and its circumference | surroundings. In related art, it is the schematic explaining the intensity | strength of the reflected light which injects into a camera along the line | wire which crosses the hollow which comprises the identification symbol, and its circumference | surroundings. It is a figure explaining the state by which wiring was formed in the vicinity of the hollow which comprises an identification symbol. It is a figure explaining the result of having measured the reflected light intensity from the surface of a semiconductor wafer which the camera of the identification symbol reader in related technology receives along the line which crosses the hollow and wiring which comprise an identification symbol. It is a figure explaining an example which an error produces in reading of an identification symbol. It is a figure explaining the flow of the process in which the image processing unit of related technology reads the character which comprises an identification symbol. It is a figure explaining the image processing performed with the image processing unit of related technology. It is a figure explaining the dictionary data with respect to character "3". It is a figure explaining the image data in which the wiring 28 is applied to the end of the character “3” constituting the identification symbol 12. It is a figure explaining the dictionary data with respect to the characters "3" and "8". It is the schematic which looked at the identification symbol reader provided with the oblique illumination unit from the side. It is a figure explaining the state which irradiated the oblique light to the hollow which comprises an identification symbol, and its circumference | surroundings.

Explanation of symbols

2 ... Identification symbol reader (with epi-illumination unit)
4 ... Illumination 6 ... Half mirror 8 ... Epi-illumination unit 10 ... Semiconductor wafer 12 ... Identification symbol 14 ... Epi-illumination light 16 ... Camera 18 ... Image processing unit 20 .. Recess 22 ... Regular reflection light 24 generated around the recess 24 ... Reflected light generated at the recess 26 ... Surrounding the recess 28 ... Wiring 30 ... Reflected light from the wiring
32 ... Flat area on the wafer surface (area around the recess where no wiring is formed)
34... Pixel 36... Pixel 38 in dictionary data... Corresponding pixel in image data 40... Pixel 42 corresponding to wiring. Device 44 ... Oblique light 46 ... Oblique light unit 48 ... Illumination unit 50 ... Camera 52 ... Control unit
54 ... Identification symbol reader 58 in the present invention 58 ... Light (low oblique light)
60 ... Low oblique light irradiated to the periphery of the recess 62 ... Low oblique light irradiated to the recess 64 ... Low oblique light reflected by the recess 66 ... Light emitting member 68 ... Rotating member 72 ...・ Axis 74 parallel to the semiconductor wafer: a depression sufficiently exposed to low oblique light

Claims (7)

Between multiple processes of manufacturing semiconductor devices on semiconductor wafers,
An identification symbol consisting of an array of a plurality of depressions marked on the semiconductor wafer,
In the identification symbol reading method for reading by irradiating light around the depression and the depression,
A first step of setting an angle formed by the traveling direction of the light with respect to the surface of the semiconductor wafer at the position of the identification symbol so that the depression is observed brighter than the surroundings;
An identification symbol reading method comprising a second step of reading the identification symbol after the first step. The identification symbol reading method according to claim 1,
The first step includes
An identification symbol reading method comprising adjusting the angle so that a coincidence ratio between dictionary data registered in advance as an image of the identification symbol and image data read as the identification symbol is equal to or greater than a predetermined value. The identification symbol reading method according to claim 2,
The first step includes
In addition to the angle, either or both of the intensity of the light on the semiconductor wafer and the position where the intensity is maximized are adjusted. In the identification symbol reading method according to any one of claims 1 to 3,
The first step includes
The identification symbol reading method, which is performed when the surface state of the recess and the region including the periphery of the recess is changed by the step. Between multiple processes of manufacturing semiconductor devices on semiconductor wafers,
An identification symbol consisting of an array of a plurality of depressions marked on the semiconductor wafer,
In the identification symbol reading device that reads by irradiating light around the depression and the depression,
An illumination unit that emits the light;
A camera that captures an area composed of the depression and the periphery of the depression;
An image processing unit for processing the image and extracting the identification symbol;
The lighting unit is
An angle formed by the traveling direction of the light with respect to the semiconductor wafer at the position of the identification symbol,
An identification symbol reading device characterized in that it can be set so that the depression is observed brighter than the front periphery. The identification symbol reader according to claim 5,
A control unit for controlling the lighting unit based on the image data;
The image processing unit includes:
Creating the image data in which one of the first value corresponding to a bright pixel and the second value corresponding to a dark pixel is associated with each pixel according to the brightness of the pixel;
The control unit is
By controlling the lighting unit,
The identification symbol reading device, wherein the angle is adjusted so that a pixel corresponding to the depression has the first value and a pixel corresponding to the surroundings has the second value. A first step of reading an identification symbol formed on an array of a plurality of depressions marked on a semiconductor wafer by irradiating light around the depressions and the depressions;
A second step of setting an angle formed by the traveling direction of the light with respect to the surface of the semiconductor wafer at the position of the identification symbol so that the depression is observed brighter than the surroundings;
A method of manufacturing a semiconductor device, comprising a third step of reading the identification symbol after the second step.

Images