JP2013004565A - Image sensor structure, manufacturing method of image sensor and image sensor - Google Patents

Abstract

Variation in film thickness of a color filter of an image sensor is reduced.
An image sensor structure includes a plurality of image sensors formed in an array on a silicon substrate 51 and a scribe line between adjacent image sensors which is cut when the plurality of image sensors are separated by dicing. Line. Each of the plurality of image sensors is formed of a plurality of photoelectric conversion elements composed of a diffusion layer 52 formed on the silicon substrate 51 and a conductor surrounding the plurality of photoelectric conversion elements along the outer periphery of each of the image sensors. Guard wiring 81 is provided. Then, a dummy wiring 82 formed in the same layer by the same process as the guard wiring 81 is formed in the scribe line portion 53.
[Selection] Figure 8

Description

  The present invention relates to an image sensor for reading image information such as a copying machine, a facsimile, and an OCR.

  In image reading devices such as a facsimile, a scanner, and a copying machine, a contact image sensor is used to read an image. The contact image sensor includes a line image sensor having a size of 10 to 20 mm in which a plurality of photoelectric conversion elements are linearly arranged in a desired reading length. The contact image sensor is a line image sensor, a light source module for irradiating light on a manuscript, a lens for collecting light reflected on the manuscript, and an electric signal photoelectrically converted and amplified by the line image sensor. It consists of a signal processing circuit for digital conversion and a housing frame for holding these components.

  In a line image sensor, various elements are formed on a silicon substrate using a CMOS (Complementary Metal Oxide Semiconductor) process, and a plurality of photodiodes (PD: Photo Diode) which are photoelectric conversion light receiving elements are linearly arranged at equal intervals. ing. In addition to the PD, the line image sensor mainly includes a signal processing unit that amplifies the output from the PD, and a latch circuit that sequentially reads sequentially in accordance with an external start signal.

  With the color of image reading devices in recent years, line image sensors have also been colored. In the color line image sensor, a color filter that transmits a specific wavelength band is mounted on the PD. The main transmission wavelength bands of the color filters are red (630 nm), green (520 nm), and blue (450 nm), which are the three primary colors. In the color line image sensor, PDs corresponding to each color filter are arranged in a line for each color on a straight line. Similarly, the signal amplifier circuit is mounted corresponding to each color PD.

  Signals output from the color line image sensor are generally converted into image information and used for computer screen display and printing with a printer. With the colorization of the read image, the variation between the pixels of the photoelectric conversion signal output from the color image sensor becomes apparent as unevenness and streaks in the image information. Reducing the photoelectric conversion and characteristic variation of the color line image sensor is a big problem.

  For example, Patent Document 1 discloses a technique for suppressing variation in the value of an electric signal output from each photoelectric conversion element of a line image sensor by making the thickness of a protective film uniform. The photoelectric conversion device of Patent Document 1 is a semiconductor substrate on which a plurality of line image sensor ICs arranged in the vertical and horizontal directions with a gap between each other is formed, in a region where the short sides of adjacent line image sensor ICs face each other, A pattern by dummy wiring is formed. After that, the photoelectric conversion device is cut (diced) along a scribe line between the line image sensor ICs to be separated into image sensor ICs.

  When dicing the semiconductor substrate on which the line image sensor is formed, it is necessary not to damage the line image sensor. For example, in the semiconductor wafer dicing method disclosed in Patent Document 2, a protective film separated from a protective film provided on a semiconductor element is provided at a scribe line planned portion, and the separated protective film portion is diced.

JP 2008-159974 A JP-A-1-196850

  There are various factors that cause variations in the photoelectric conversion characteristics of the PDs mounted on the color line image sensor, but they can be roughly classified into two.

  First, manufacturing variations in the semiconductor process for forming the image sensor and variations in the film thickness of the transparent film (protective film) on the photoelectric conversion element due to layout dependency occur in the image sensor chip. Variations in the intensity of light interference occur due to variations in the thickness of the transparent film. Therefore, when the light incident on the image sensor reaches the PD, the light intensity varies, and as a result, the output after the photoelectric conversion is varied.

  Second, variation in the transmission characteristics of the color filter is caused by variations in the film thickness of the color filter disposed on the photoelectric conversion element. When the light incident on the image sensor reaches the PD due to the variation in the film thickness of the color filter, the intensity of the light is generated, and the output after the photoelectric conversion varies.

  The above two phenomena often occur in the photoelectric conversion light receiving element in the vicinity of the end in the longitudinal direction of the line image sensor. The photoelectric conversion device of Patent Document 1 is one of means for improving the first problem. According to Patent Document 1, when the source gas is generated by the plasma CVD method, the source gas is uniformly deposited not only on the line image sensor IC but also on the dummy wiring. For this reason, it is described that a protective film (transparent film) having a uniform film thickness is formed on the line image sensor IC.

  In general, the configuration of each layer formed in the line image sensor differs between the photoelectric conversion element portion and the wiring portion, so that the surface of the protective film of the image sensor is uneven. The unevenness of the protective film is one of the causes of variations in the film thickness of the color filter. With the technique of Patent Document 1, the unevenness on the surface of the protective film cannot be eliminated. The variation in film thickness of the second color filter cannot be solved by Patent Document 1 and Patent Document 2.

  The present invention has been made in view of the above-described circumstances, and an object thereof is to reduce the variation in the film thickness of the color filter of the image sensor.

  In order to achieve the above object, an image sensor structure according to the present invention includes a plurality of image sensors formed in an array on a substrate, and adjacent images that are cut when the substrate is diced to separate the image sensors. An image sensor structure including a scribe line between the sensors, each of the image sensors surrounding a plurality of photoelectric conversion elements formed on a substrate and the plurality of photoelectric conversion elements. Guard wiring formed of a conductor along each outer periphery. The scribe line includes dummy wirings formed in the same layer along the scribe line by the same process as the guard wiring.

  According to the present invention, the unevenness generated in the scribe line portion of the substrate on which the plurality of image sensors are formed can be reduced, and as a result, the variation in the film thickness of the color filter of the image sensor can be reduced.

It is a figure which shows the image sensor structure in which the several image sensor was formed in the arrangement | sequence on the silicon substrate. It is a partial enlarged view of an image sensor structure. It is a block diagram which shows the example of the circuit structure of an image sensor. It is sectional drawing of the edge part and scribe line of an image sensor before forming a color filter. It is sectional drawing of the image sensor in which the color filter was formed. It is sectional drawing of the image sensor edge part and scribe line which do not use CMP construction method. It is a top view which shows a part of image sensor structure which concerns on Embodiment 1 of this invention. 2 is a cross-sectional view of an end portion of an image sensor and a scribe line according to Embodiment 1. FIG. 6 is a diagram showing a manufacturing process of guard wiring and dummy wiring of the image sensor structure according to Embodiment 1. FIG. 2 is a cross-sectional view of an image sensor in which a color filter according to Embodiment 1 is formed. FIG. 4 is a cross-sectional view of an end portion of an image sensor and a scribe line according to a different example of the first embodiment. FIG. It is sectional drawing of the edge part and scribe line of the image sensor which concerns on Embodiment 2 of this invention. It is sectional drawing of the image sensor which formed the slit in the protective film of the both sides of a scribe line.

  In the color filter manufacturing process, in order to make the film thickness of the color filter uniform, the base portion on which the color filter is to be mounted is covered with a transparent resin for flattening. The base described here refers to the uppermost protective film surface of the image sensor formed in the semiconductor manufacturing process. If the surface irregularities of the protective film are large, the flattening process must be performed a plurality of times without being flattened by applying the transparent resin once. When the planarization step is performed a plurality of times, the number of steps and the amount of transparent resin increase, leading to an increase in cost. In particular, the unevenness on the surface of the protective film is conspicuously shown in a scribe line portion provided for separating and cutting the image sensors arranged at equal intervals on the silicon substrate. The embodiment of the present invention provides means for reducing the unevenness generated in the scribe line portion, and reduces the burden on the flattening process in the color filter process.

(Embodiment 1)
In order to facilitate understanding, a general line image sensor will be described first. Hereinafter, the line image sensor is simply referred to as an image sensor. FIG. 1 is a view showing an image sensor structure formed by arranging a plurality of image sensors on a silicon substrate. In the image sensor structure 1, the image sensors 2 are formed on a circular silicon substrate 51 at an equal interval.

  The image sensor 2 is generally manufactured by a CMOS process, and its chip size is a rectangle having a lateral width W of about 10 to 20 mm in the X direction and a longitudinal width L of about 0.4 to 0.8 mm in the Y direction. In the image sensor 2, photodiodes (hereinafter referred to as PDs) having a constant area, which are photoelectric conversion elements, are arranged at a constant density and interval in the X direction. The density / interval of the PD is, for example, 600 dpi pitch.

  FIG. 2 is a partially enlarged view of the image sensor structure. Between the image sensors 2 formed at equal intervals, a scribe line 21X sandwiched between the long sides of the image sensor 2 and extending in the X direction, and a scribe line 21Y sandwiched between the short sides of the image sensor 2 and extending in the Y direction. Is provided. The scribe lines 21X and 21Y are regions for cutting the silicon substrate 51 in order to separate the plurality of image sensors 2 one by one. The scribe line 21X and the scribe line 21Y are formed in a lattice shape, and the periphery of each image sensor 2 arranged in the XY direction is surrounded by the scribe lines 21X and 21Y. Both the width SD of the scribe line 21X and the width SW of the scribe line 21Y have equal intervals or different intervals. The width SD and the width SW are approximately 50 to 200 μm.

  Each of the image sensors 2 includes pixels 22R, 22G, and 22B that detect light of three colors. Each of the pixels 22R, 22G, and 22B includes a PD and a color filter. The pixels 22R, 22G, and 22B are arranged on a straight line in one row in the X direction. In FIG. 2, only about 30 pixels 22R, 22G, and 22B are drawn by one image sensor 2, but in actuality, several hundred or more pixels are formed in the X direction. The image sensor 2 is formed with a wire bonding pad 23 for connecting wiring for supplying power and taking out a signal.

  FIG. 3 is a block diagram illustrating an example of a circuit configuration of the image sensor. The circuit of the image sensor 2 includes a pixel circuit 34, a selector 33, a start signal control unit 31, and an amplifier circuit 32. In the pixel circuit 34, a plurality of transistors 37 and an analog memory 35 that amplify a signal from the PD 36 are arranged at the subsequent stage of the PD 36 of the image sensor 2. The selector 33 sequentially reads out the signals of the plurality of pixel circuits 34 and serializes them into serial signals. The start signal control unit 31 is a circuit for controlling the sequential reading of the selector 33 by an external start signal. The serialized signal (analog signal) of the pixel circuit 34 is amplified by the amplifier circuit 32 and output to the outside. In the color image sensor, the circuit configuration shown in FIG. 3 is required independently for each of the three colors red, green, and blue, and three circuits of the above-described configuration are arranged. Next, a cross-sectional configuration of the image sensor 2 will be described.

  FIG. 4 is a cross-sectional view of the end portion of the image sensor and the scribe line before forming the color filter. FIG. 4 shows a section taken along line AA in FIG. FIG. 4 shows the end portion 58 of the image sensor 2 on both sides, and the cross section (scribe line portion 53) of the scribe line 21Y in the center. A substrate used in the CMOS process is generally a P-type silicon substrate 51. The PD 36 which is a photoelectric conversion element is formed on the silicon substrate 51 as an N-type diffusion layer 52 in which arsenic or phosphorus is ion-implanted. The P-type silicon substrate 51 and the N-type diffusion layer 52 are used as PN junction diodes to form a PD 36 that is a photoelectric conversion element.

  An insulating film 54a is disposed on the silicon substrate 51 on which the diffusion layer 52 is formed. A wiring layer 55a formed of a conductive material such as aluminum for electrically connecting a circuit of a separately formed transistor or the like (not shown) is disposed on the insulating film 54a. Although not shown in FIG. 4, a contact for connecting the wiring layer 55a and the photoelectric conversion element is formed. Further, in order to configure the circuit of the image sensor 2, insulating films 54 b and 54 c and contacts (also referred to as vias, not shown) connecting the wiring layers 55 b and 55 c and the wiring layers are formed on the wiring layer 55 a. .

  In the vicinity of the PD 36 (diffusion layer 52), the wiring layers 55a, 55b, and 55c are arranged as a light shielding layer for optically separating a plurality of arranged PDs 36. In FIG. 4, wiring layers 55a, 55b, and 55c as light shielding layers are shown. These light shielding layers are generally arranged with the maximum number of layers that can be manufactured in a semiconductor process for manufacturing a line image sensor. For example, if three wiring layers can be formed in a semiconductor process, the number of light shielding layers is three. However, this is not necessarily the case. Since the thickness of the layer to be formed differs depending on the pattern of the wiring layers 55a, 55b, and 55c, unevenness 57 is formed on the surface of the protective film 56a.

  In a recent miniaturized semiconductor process, the wiring layer is mainly a multilayer. In order to stack the multilayer wiring layers with high accuracy, the insulating layer disposed between the wiring layers is planarized using a method called chemical mechanical planarization (CMP), and then the upper wiring layer is formed. . In the semiconductor process using the CMP method, since the thickness of each layer is controlled and laminated, the variation in film thickness from the silicon substrate 51 to the protective film surface which is the uppermost surface of the image sensor 2 is reduced.

  After forming the uppermost protective film 56a, a color filter is formed. FIG. 5 is a cross-sectional view of an image sensor in which a color filter is formed. In the color filter forming step, the color filter film is formed on the surface of the protective film 56a by using a spin coat method. For this reason, if there are irregularities on the surface of the protective film 56a, the thickness of the color filter is adversely affected. Therefore, in order to reduce unevenness on the surface of the protective film 56a, a transparent resin is applied on the surface of the protective film 56a and flattened to form a transparent flattened film 62 (flattening step).

  Since the color filter film is formed by spin coating, it is affected by the unevenness of the protective film on the entire silicon substrate. Therefore, the planarization process is performed on the protective film on the entire silicon substrate. In particular, the scribe line portion 53 disposed between the line image sensors is a portion where the unevenness is likely to occur, and is a target for flattening.

  In the case of a monochrome line image sensor in which the color filter 61 is not formed, the protective film 56b (see FIG. 4) of the scribe line portion 53 is generally removed by etching. This is because when the image sensor 2 is separated and cut, the protective film 56a is not cracked or wrinkled. In that case, a difference occurs between the thickness H from the silicon substrate 51 of the image sensor 2 to the surface of the protective film 56a and the thickness U of the scribe line portion 53, resulting in a deep uneven state. There is a restriction on the amount of resin to be applied in one flattening step, and when the unevenness amount is large, the flattening step must be performed a plurality of times without being flattened by one resin application. Therefore, in order to reduce the number of planarization steps, it is desirable not to remove the protective film 56b of the scribe line portion 53.

  After performing the planarization step, a color filter film is formed by a spin coating method, and unnecessary portions are removed by etching or the like, and the color filter 61 is formed only in necessary portions such as the upper portion of the PD 36. This series of steps is performed on the color filters 61 for three colors of red, green, and blue. Since FIG. 5 is a cross section parallel to the arrangement of the pixels 22R, 22G, and 22B for each color, only one color filter 61 is shown. Thereafter, a color filter protective film 63 for protecting the color filter 61 is formed, and unnecessary films 64 on the scribe line portion 53 and the wire bonding pad 23 are removed, thereby completing the color filter process.

  By using the CMP method in the semiconductor manufacturing process described above, a uniform film thickness up to the protective film 56a is ensured, but the manufacturing cost tends to increase because the manufacturing man-hour is increased. In recent years, an image sensor that requires a low price often uses a semiconductor manufacturing process that does not use the CMP method in order to reduce the manufacturing cost. In this case, unlike the semiconductor manufacturing process using the CMP method, as shown in FIG. 6, many irregularities 72 depending on the wiring layout arrangement occur on the surface of the protective film 56a. The same applies to the scribe line portion 53, and a problem arises that the unevenness 71 having a large depth D is generated as compared with the case where the CMP method is used.

  Although the unevenness 71 and the unevenness 72 have the same depth, the unevenness 71 of the scribe line portion 53 has a larger area than the unevenness 72 in the image sensor 2 and is continuous with the outer periphery of the image sensor 2. The influence of is great. As the portion near the outer periphery of the image sensor 2 is flattened, the thickness of the color filter 61 varies greatly.

  FIG. 7 is a plan view showing a part of the image sensor structure according to Embodiment 1 of the present invention. In the image sensor structure 1 of the first embodiment, the dummy wirings 41X and 41Y are formed on the scribe lines 21X and 21Y. FIG. 8 is a cross-sectional view of the end portion of the image sensor and the scribe line according to the first embodiment. FIG. 8 shows a cross section taken along line BB in FIG.

  A guard wiring 81 for preventing moisture from entering the inside of the image sensor 2 from the outside is disposed at the end 58 of the image sensor 2. Since the guard wiring 81 is formed along the outer periphery of the image sensor 2, it is also called a guard ring or a seal ring. The guard wiring 81 is the maximum number of wiring layers possible in the semiconductor manufacturing process, and is composed of wiring layers formed of conductors and contacts (also referred to as vias) that connect the wiring layers. The thickness from the silicon substrate 51 to the protective film 56 a on the guard wiring 81 is the largest in the image sensor 2. The scribe line portion 53 is adjacent to the guard wiring 81, and normally, no wiring layer is disposed on the scribe line portion 53 as shown in FIG. Therefore, the scribe line portion 53 is a portion where the thickness becomes the smallest and the unevenness increases even when the protective film 56b (FIG. 4) is not etched.

  For the purpose of reducing the unevenness caused by the film thickness difference between the guard wiring 81 and the scribe line portion 53, changing the layer configuration of the guard wiring 81 loses the function of preventing moisture absorption and improves the reliability of the line image sensor. It is difficult because there is a problem of lowering. However, it is easy to change the layer structure of the scribe line portion 53 without affecting the reliability problem.

  When the CMP method is not used, the thickness to the surface of the silicon substrate 51 and the protective film 56a depends on the wiring pattern arranged in each layer. The portion where the wiring layers 55a, 55b and 55c are not disposed, for example, the upper portion of the PD 36 is thin, and the portion where all the wiring layers are disposed is thickest.

  In the first embodiment, the scribe line portion 53 is provided with a dummy wiring 82 composed of a wiring layer 82a and a contact 82b. The dummy wiring 82 in FIG. 8 is a cross section of the dummy wiring 41Y shown in FIG. The dummy wirings 82 (dummy wirings 41X and 41Y) are formed in the same layer in the same process as the guard wiring 81 along the scribe lines 21X and 21Y. By providing the dummy wiring 82 in the scribe line portion 53, the layer configuration of the card layer 81 and the scribe line portion 53 are the same, and the thickness of the guard wiring 81 portion and the scribe line portion 53 to the protective film 56a is substantially equal. .

  By making the layer configuration of the guard wiring 81 and the dummy wiring 82 the same, the deep unevenness 71 is not formed in the guard wiring 81 portion and the scribe line portion 53. Since the guard wiring 81 has a wiring layer configuration with the maximum number of layers of the image sensor 2, the dummy wiring 82 has the layer configuration of the guard wiring 81 unless a new film formation process that is not included in the process of forming the image sensor 2 is added. When the layer structure is different, the thickness of the scribe line portion 53 is thinner than the thickness of the guard ring portion. That is, when the guard wiring 81 and the dummy wiring 82 have the same layer structure, the depth of the unevenness of the guard wiring 81 and the scribe line portion 53 is minimized.

  The guard wiring 81 and the dummy wiring 82 are formed as follows. As described with reference to FIG. 4, arsenic or phosphorus ions are implanted into a P-type silicon substrate 51 to form an N-type diffusion layer 52. Further, a CMOS transistor (not shown) for forming a signal processing circuit is formed in each region. An insulating layer 54a made of, for example, an oxide film is formed thereon.

  FIG. 9 is a diagram illustrating a manufacturing process of guard wiring and dummy wiring of the image sensor structure according to the first embodiment. In order to electrically connect wiring to the photoelectric conversion element and the CMOS transistor, a contact hole is formed in the insulating layer 54a by etching or the like. At this time, as shown in FIG. 9A, a contact hole 54h is also formed in the guard wiring 81 and the dummy wiring 82.

  Next, an aluminum film 55f as a conductive material is deposited (see FIG. 9B), and a wiring layer 55a is formed with a predetermined wiring pattern by etching (see FIG. 9C). At this time, simultaneously with the wiring pattern, the aluminum film 55f in the portions of the guard wiring 81 and the dummy wiring 82 is left, and these patterns are formed. In addition, a light shielding layer is formed simultaneously with the wiring pattern. Then, an insulating film 54b is formed on them (see FIG. 9D). 9A to 9D, contact holes are formed in the insulating film 54b, and the wiring layer 55b is formed on the insulating film 54b. At the same time, the guard wiring 81 is formed in the same layer as the wiring layer 55b. Then, the dummy wiring 82 and the light shielding layer are formed. Further, the insulating film 54c, the wiring layer 55c, the guard wiring 81, the dummy wiring 82, and the light shielding layer which are the same layer as the wiring layer 55c are formed.

  FIG. 10 is a cross-sectional view of the image sensor in which the color filter according to Embodiment 1 is formed. After the wiring layers 55a, 55b and 55c (including the light shielding layer), the guard wiring 81 and the dummy layer 82 are formed, a transparent protective film 56 made of, for example, silicon nitride (SiN) is formed thereon. After forming the protective film 56 covering the guard wiring 81 and the dummy wiring 82, the transparent flattening film 62, the color filter 61, and the color filter protective film 63 are formed as in FIG.

  As described above, the manufactured image structure 1 is diced and cut along the scribe lines 21X and 21Y (scribe line portion 53), thereby separating the image sensors 2 one by one. The separated image sensor 2 is fixed to the housing frame and wired between the wire bonding pad 23 and the signal processing circuit or the external terminal. Further, a light source module and a lens are incorporated and sealed to complete a contact image sensor.

  As described with reference to FIG. 6, the unevenness 71 (see FIG. 6) of the scribe line portion 53 that has a large area and has a large influence on the flattening process is eliminated. And the variation in the film thickness of the color filter 61 is reduced.

  FIG. 11 is a cross-sectional view of an end portion of an image sensor and a scribe line according to a different example of the first embodiment. In the example shown in FIG. 11, in order to laminate the multilayer wiring layers with high accuracy, the upper wiring layer is formed after the insulating layer disposed between the wiring layers is planarized using the CMP method. In the semiconductor process using the CMP method, since the thickness of each layer is controlled and laminated, the variation in film thickness from the silicon substrate 51 to the surface of the protective film 56 which is the uppermost surface of the image sensor 2 is reduced.

  Also in the example of FIG. 11, the dummy wiring 82 is formed in the same layer in the scribe line portion 53 by the same process as the guard wiring 81. By arranging the dummy wiring 82, the unevenness of the scribe line portion 53 can be made extremely small. Since the insulating layer disposed between the wiring layers is planarized using the CMP method, the unevenness inside the image sensor 2 is reduced as compared with the configuration of FIG. As a result, the accuracy of planarization is further high, and the number of planarizations can be reduced. And the variation in the film thickness of the color filter 61 can be made smaller.

  In the image sensor structure 1 according to the first embodiment, even when a semiconductor process that does not use the CMP method is used, unevenness is unlikely to occur in the scribe line portion 53. The unevenness on the surface of the protective film 56 of the scribe line portion 53 is not limited to the portion adjacent to the end portion 58 of the image sensor 2. In any place where the guard wiring 81 on the outer periphery of the image sensor 2 is disposed, the effect of reducing the unevenness on the surface of the protective film 56 of the scribe line portion 53 is obtained.

  By adopting a configuration in which the dummy wiring 82 is disposed in the scribe line portion 53, the unevenness on the surface of the protective film 56 of the scribe line portion 53 is reduced. Since the unevenness on the surface of the protective film 56 is reduced, the surface can be surely flattened before the color filter film is formed by the flattening step. As a result, the thickness of the color filter film can be made uniform, and variations in the film thickness of the color filter 61 of the image sensor 2 can be reduced. The dummy wiring 82 can reduce the load of the flattening process in the color filter manufacturing process. Accordingly, it is possible to provide the image sensor 2 having a color filter characteristic with little variation at low cost.

(Embodiment 2)
In the image sensor structure 1 according to the second embodiment, before dicing the image sensor 2, grooves are formed along the scribe lines 21X and 21Y in the protective film 56 around each image sensor 2.

  As shown in FIG. 8, the image sensor structure 1 according to the first embodiment is arranged around the image sensor and the protective film 56 b of the scribe line portion 53 in the configuration having the dummy wiring 82 in the scribe line portion 53. The protective film 56a on the guard wiring 81 is connected. As described above, in the silicon substrate 51 on which a semiconductor IC including a general image sensor is formed, the protective film 56b of the scribe line portion 53 is removed before dicing. This is because, when the separation / cutting is performed by a dicing apparatus using a diamond cutter, the rotation of the diamond cutter does not cause the protective film 56a to crack or bend. When the protective film 56a is cracked or wrinkled, the wiring layer at the end of the guard wiring 81 is exposed, and the function of the guard wiring 81 that prevents moisture from entering, which is the original purpose, is impaired.

  FIG. 12 is a cross-sectional view of an end portion of an image sensor and a scribe line according to Embodiment 2 of the present invention. As in the first embodiment, after the guard wiring 81 and the dummy wiring 82 are formed, the protective film 56 that covers the guard wiring 81 and the dummy wiring 82 is formed. Then, a groove 91 is formed in the protective film 56 between the guard wiring 81 and the dummy wiring 82. The protective film 56 is formed by depositing silicon nitride (SiN) or the like by plasma CVD, for example. Then, for example, the groove 91 is formed by etching.

  The grooves 91 are continuously formed around the respective image sensors 2 along the scribe lines 21X and 21Y. Since the groove 91 is formed around each of the image sensors 2, it is formed on both sides of the scribe line portion 53. The depth of the groove 91 reaches at least from the surface of the protective film 56 to the layer below the protective film 56. That is, the protective film 56 is divided by the groove 91.

  When cracks or wrinkles occur during dicing by forming the grooves 91 in the protective film 56, the protective film 56 is divided by the grooves 91, so that the progress of the cracks and wrinkles is stopped before reaching the guard wiring 81. Can do.

  By adopting such a configuration, it is possible to provide a color line image sensor having a color filter characteristic with little variation at a low cost in which the generation of cracks and wrinkles in the protective film 56 is suppressed.

1 Image sensor structure
2 Image sensor
21X, 21Y scribe line 22R, 22G, 22B pixels
23 Wire bonding pad
36 Photodiode (PD)
41X, 41Y dummy wiring
51 Silicon substrate
52 Diffusion layer
53 Insulating film 54 scribe line part 54a, 54b, 54c
54h Contact hole 55a, 55b, 55c Wiring layer
55f Aluminum film 56, 56a, 56b Protective film
58 End of image sensor
61 Color filter
62 Transparent planarization film
63 Color filter protective film
81 Guard wiring
82 Dummy wiring
91 groove

Claims (5)

An image sensor comprising: a plurality of image sensors formed in an array on a substrate; and a scribe line between adjacent image sensors that is cut when the substrate is diced to separate the image sensors. A structure,
Each of the image sensors
A plurality of photoelectric conversion elements formed on the substrate;
A guard wiring that surrounds the plurality of photoelectric conversion elements and is formed of a conductor along the outer periphery of each of the image sensors;
The image sensor structure, wherein the scribe line includes a dummy wiring formed in the same layer in the same process as the guard wiring along the scribe line. A protective layer covering the guard wiring and the dummy wiring;
2. The image sensor structure according to claim 1, wherein a groove is formed around each of the image sensors along the scribe line in the protective layer from the surface of the protective layer to a layer below the protective layer. . A method of manufacturing an image sensor, wherein a plurality of image sensors are formed in an array on a substrate, and the image sensors are separated by cutting along a scribe line between adjacent image sensors,
An element forming step of forming a plurality of photoelectric conversion elements in each of the image sensors;
Surrounding the plurality of photoelectric conversion elements of each of the image sensors, the guard wiring of the conductor along the outer periphery of each of the image sensors, the scribe line part along the scribe line, and the same layer as the guard wiring A planarization step for forming dummy wirings in the process of
A dicing step of cutting the substrate along the scribe line to separate the image sensor;
An image sensor manufacturing method comprising: After the planarization step and before the dicing step,
A protective layer step of forming a protective layer covering the plurality of image sensors and the scribe line;
Forming a groove reaching the layer below the protective layer from the surface of the protective layer along the scribe line around each of the image sensors in the protective layer;
An image sensor manufacturing method according to claim 3.   The image sensor structure of Claim 1 or 2 is cut | disconnected by the said scribe line, and the image sensor isolate | separated.

Images