JP2009295739A - Semiconductor image sensor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a sensor structure of a semiconductor image sensor, which avoids poor connection and the reduction in reliability of connection, which are caused because a rear surface of a thinned semiconductor substrate is not flat, and furthermore avoids risks of destruction of the semiconductor substrate due to a sudden force apt to occur after a thinning step, and to provide a method for manufacturing the semiconductor image sensor. <P>SOLUTION: The semiconductor image sensor is most principally characterized in that a surface of a microlens group for condensation is brought into contact with a lower surface of a cover glass 31 in the thinning step of a semiconductor substrate 2 in order to prevent a phenomenon that a principal region of the image sensor where the microlens group 30 etc. are arranged, is pushed toward the cover glass. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a low profile and small semiconductor image sensor.

Image sensors that utilize semiconductor technology have become increasingly widespread in many image input devices due to advances in the number of pixels and miniaturization that have led to advances in semiconductor technology. In particular, digital cameras and mobile phones are increasingly used as image recording and scanners, and there is a strong demand for downsizing and further lowering the price of image sensors, including further increasing the number of pixels and reducing the height.
The semiconductor image sensors currently in practical use are mainly CMOS type and CCD type, and are assembled by well-known semiconductor technologies such as chip separation, wire bonding, and package encapsulation. In the sensor, a photosensitive region, a signal readout region, a signal processing / driving region, and the like are integrated on one main surface of the Si semiconductor, and wire bonding is provided for electrical connection from the main surface side to an external circuit.
In a stationary image input apparatus, if a semiconductor image sensor having the structure having the number of pixels satisfying the resolution characteristics required by the apparatus is mounted, a practical apparatus can be realized.

  However, in portable devices typified by mobile phones, there is a strong demand for miniaturization of semiconductor image sensors, and the conventional structure cannot cope with them. In particular, it has been a challenge to reduce the thickness of the sensor, that is, to reduce the height and to make the package size as close as possible to the chip size. FIG. 24 is a structural example of a semiconductor image sensor configured as a color image sensor in order to satisfy such a requirement. Reference numeral 1 denotes a sensor portion manufactured by a well-known semiconductor process, a semiconductor substrate 2 made of Si or the like, It is composed of a plurality of photodiodes 3 which are photosensitive elements provided in the vicinity of the surface, a filter 4 for color separation, a microlens group 5 for enhancing the light collection effect, and the like. 6 is a cover glass which is a surface protection function, and is bonded to 1 by an adhesive layer 7. The wiring layer for reading signals from the photodiode includes a pad region 8 for connection to an external circuit, and electrical connection to the region is formed with a through hole from the back side of the substrate 2, This is realized on the back side of the semiconductor substrate 2 by the filled conductive material 9 and the connection terminal group 10 in the form of a ball grid array. Such a structure is advantageous in that it can be mounted in an area substantially equal to the area of the image sensor and can be reduced in height as compared with the connection means from the surface side of the semiconductor substrate 2 using a bonding wire that has been widely used conventionally. Such advantages can greatly contribute to miniaturization and thinning of mobile phones and cameras that are application devices of image sensors.

  In FIG. 24, it is desirable to reduce the thickness of the substrate 2 in order to facilitate the through hole processing and further reduce the height. This thinning process is performed when the semiconductor process is completed or when the cover glass is provided. After such a thinning process is completed, processing of the through hole and filling of the 9 are performed.

  FIG. 25 is a diagram illustrating the semiconductor substrate thinning process in detail. This figure shows a state in which the four structures shown in FIG. 24 are arranged in a line, and the alternate long and short dash line 20 is the boundary between the structures. In the figure, the same numbers as those in FIG. 24 indicate the same components. In FIG. 2A, the thickness of the substrate 2 is approximately several hundred μm although it depends on the diameter of the wafer used in the semiconductor process. In FIG. 5B, the substrate is thinned and processed into a thin substrate 21. The thinning step is performed by a well-known technique called back grinding, in which a force is applied from the two back surfaces and mechanically and chemically removes the back surface. In this technique, the cover glass surface is protected by a flat plate made of resin or the like, and at the same time the strength of the entire wafer is maintained, but this is not shown in the figure. In the thinning process, the substrate 2 is processed to a thickness of about 100 μm or less. However, as the thickness decreases, a phenomenon occurs in which the main region of the image sensor in which the microlens group, the color filter, and the like are arranged is pushed into the cover glass side. This phenomenon is caused by the existence of a space between the cover glass 6 and the sensor portion 1. Such a phenomenon is shown in FIG. 2C, and becomes prominent when the thickness of 21 is about several tens of μm. When the thinning process is completed, the force applied to 21 is removed, so that the distance between the main area of the image sensor and the cover glass returns to the distance before the back surface polishing process. As a result, as shown in FIG. 4D, the rear surface side of 21 is not flat, and the main region is in a convex state. That is, unevenness occurs on the back surface side of 21, and the state that should originally be flat is not maintained, and many inconveniences occur in the subsequent assembly process. As an example of such inconvenience, since the surface of the through hole and the filled conductive material is not flat, the surface of the ball grid array-like connection terminal group that is a connection means with an external circuit is not flat, and the connection is poor. Generation of the network and a decrease in connection reliability. Furthermore, even when these inconveniences are not present, there is an increased risk that the 21 will be destroyed by the force suddenly applied from the back side of the 21 in the assembly process. In order to reduce this risk, it is conceivable to attach a reinforcing flat plate made of glass or the like on the back side of 21. Disadvantageous.

  Such a problem is caused by the existence of a space between the main region of the semiconductor image sensor and the cover glass. The backside thinning process in the semiconductor field is performed by forming a resin film using a resin such as polyimide or epoxy on the surface of a semiconductor substrate, as proposed in Patent Document 1 and Patent Document 2 below. After curing, the cured resin film is used as a protective reinforcement film, and a substrate is ground after the surface is affixed with a flexible protective film (BG tape). In Patent Document 1 and Patent Document 2 described below, application to a semiconductor image sensor is not necessarily mentioned, but the back surface is thinned after the surface of the semiconductor substrate is planarized.

When these methods are applied to a semiconductor image sensor equipped with a microlens, the space between the main region and the cover glass can be eliminated due to the flat shape of the surface, and there is no back surface unevenness at the end of substrate thinning. .
Japanese Patent Laid-Open No. 11-150090 JP 2005-191508 A

  However, when the backside polishing method of the above-mentioned patent document is adopted, a thermoset resin film remains on the upper surface of the microlens, making it difficult to exhibit the original optical function of the microlens. An important issue arises. Further, it is difficult to completely remove the thermosetting resin film without exposing the microlens shape to expose the surface of the microlens, and even if possible, the manufacturing process is complicated.

  The problem to be solved by the present invention is that when thinning a semiconductor image sensor on which a microlens is mounted, there is no need to deposit and remove a thermosetting resin film on the surface, which is a factor that hinders the optical function of the microlens Is to eliminate. In addition, the backside of the thinned semiconductor substrate is not flat, resulting in poor connection and reduced connection reliability. Furthermore, there is a risk of destruction of the semiconductor substrate due to sudden forces that tend to occur after the thinning process. is there.

  The present invention is a phenomenon in which the main region of a semiconductor image sensor in which microlens groups and the like are arranged is pushed into a flat plate side having at least the characteristic that it is transparent in a specific light wavelength region in a thinning process of a semiconductor substrate In order to prevent this, the surface of the microlens group for condensing is brought into contact with the flat plate and flattened.

  In addition to the photosensitive region where the pixels are arranged, a plurality of structures having the same shape as the microlens and having a similar shape or a structure having a planar extension are disposed.

  The said structure is comprised from multiple types from which a planar dimension differs.

  According to the present invention, since the microlens group is in contact with the cover glass in the thinning process of the semiconductor substrate, the main region of the semiconductor image sensor can be prevented from being pushed into the cover glass side, and at the end of the thinning process. The back surface of the semiconductor substrate becomes flat, and further, the semiconductor substrate is not destroyed even if a sudden force is applied from the back surface side.

  Since the contact is achieved at the stage of mounting the cover glass or during the thinning process of the semiconductor substrate, the flatness of the back surface of the semiconductor substrate can be easily maintained and the risk of destruction can be avoided.

  Since the contact state and the shape of the microlens are controlled by a structure disposed outside the photosensitive region in which the pixels are arranged, the uniformity and reproducibility of the shape can be easily ensured.

  The state of contact and the shape of the microlens can be controlled by the shape, type, number, and the like of the structure.

  The purpose of reducing the height and downsizing of semiconductor image sensors is without the occurrence of poor connection or lowering of connection reliability, and without the complexity of the process of providing a reinforcing plate on the back of the semiconductor substrate. Realized. Hereinafter, the present invention will be clarified by describing specific embodiments of the image sensor.

<First image sensor configuration>
FIG. 1 is a diagram in which the present invention is applied to a color image sensor, and the same reference numerals as those in FIG. 24 denote the same components. In the figure, 30 is a condensing microlens group, 31 is a cover glass, and 32 is an adhesive layer for bonding the 1 and 31 together. The biggest difference between FIG. 24 and FIG. 24 described above is that the surface of the microlens 30 is in contact with the lower surface of the cover glass. In such a structure, the topmost part of the microlens is flattened by the glass, but a generally arcuate shape other than the topmost part is maintained. Unlike an imaging lens used for camera photography or the like, the original function of a microlens is to efficiently concentrate incident light on a photodiode constituting a pixel. Since the pixel of the image sensor is composed of a photodiode which is a photosensitive region and an electronic circuit for signal readout, the area of the photodiode is smaller than the pixel area. On the other hand, incident light is irradiated to the entire pixel surface, but incident light on the electronic circuit region does not contribute to photoelectric conversion, and only light incident on the photodiode region contributes to photoelectric conversion. As a result, only a part of the light that has passed through the imaging lens is used, resulting in a decrease in photosensitivity. The microlens has the purpose of concentrating light incident on the entire surface of the pixel in the photodiode region and preventing deterioration of photosensitivity. For this purpose, the surface shape of the microlens does not need to be an arc shape, and various aberrations and optical distortions that become performance indexes in a general lens system are not important. FIG. 2 conceptually shows various cross-sectional shapes of the microlens. FIG. 2A shows an example in which the entire surface of the lens is formed by an arcuate curve 40. FIG. Such a shape can be formed into an arc shape by surface tension by applying and patterning a resin constituting the lens and then placing it in a high temperature atmosphere to soften the resin. When the microlens is an inorganic material such as Si nitride, the patterned photoresist is softened on the surface of the material to form a lens shape, and then has the same processing speed in the horizontal and vertical directions. It is known that the photoresist surface shape can be transferred to the inorganic material by a technique such as ion reactive etching. In conventional image sensors, microlenses having the shape shown in FIG. FIG. 2B shows an example in which the surface of the central portion 41 of the microlens is flat and the peripheral portion 42 forms an arcuate curve. In such a structure, if the upper part of the photodiode located at the center of the pixel is located in the flat part, and the curved part is located in the upper part of the electronic circuit area around the pixel that does not inherently have photosensitivity. The light collecting function of the microlens is maintained. The shape shown in FIG. 5B can be realized by controlling the processing conditions such as the softening time and the atmospheric temperature using the above-described method of softening the resin. However, in the image sensor configuration shown in FIG. 1, the shape shown in FIG. 2 (b) is realized by pressing the cover glass, which is essentially different from realizing the shape by controlling the processing conditions. Yes.

  In FIG. 1, a cover glass 31 is bonded to the sensor portion 1 with an adhesive layer 32. The thickness of the adhesive layer can be selected as appropriate, provided that the topmost part of the microlens 30 comes into contact with the lower surface of the cover glass 31 due to shrinkage accompanying solidification of the adhesive layer. As an example, an adhesive layer having a thickness approximately equal to the thickness of the microlens before the deformation (the distance from the lens bottom surface to the topmost portion) is provided, and the cover is reduced by reducing the thickness accompanying solidification of the adhesive layer. For example, the lower surface of the glass crushes and deforms the topmost part of the microlens.

  In FIG. 1, the thickness of the cover glass 31 can be appropriately set, but is preferably several hundred μm or more. This numerical range is to prevent the central portion of the 31 from being deformed upward by the reaction force and repulsive force of the 30 when the topmost part of the microlens group 30 is deformed via the adhesive layer 32. Is set. When the upper center portion of the 31 is deformed, the cross-sectional shape of each microlens is not uniform, and the flat portion at the top of the microlens becomes smaller near the center of the sensor portion, and near the periphery of the sensor portion. Then, the flat portion becomes large. Such a variation in cross-sectional shape causes a variation in the incident state of light on each photodiode, and induces a problem that uniformity of photosensitivity is not ensured.

  FIG. 1 shows a case where the microlens is made of resin and is composed of a single convex lens. However, in the present invention, there is no limitation on the configuration of the microlens. For example, it may be composed of a first lens made of an inorganic substance between a photodiode and a color filter, a color filter on the lens, and a second convex lens made of resin on the color filter. In such a two-group two-lens microlens system, silicon nitride or the like is well known as the first lens material. Further, the first lens shape is not necessarily a convex lens shape. For example, by selecting the material of the lens and the refractive index of the surrounding material surrounding the lens, a desired light condensing effect can be obtained even with a concave lens shape. FIG. 1 shows an image sensor configuration in which a color filter is mounted for color imaging. However, color separation is unnecessary in the black-and-white image sensor, and it is not necessary to mount the color filter. This image sensor form includes a configuration in which such a color filter is not mounted.

  FIG. 3 shows a process of thinning the semiconductor substrate with the structure of FIG. In the figure, the same reference numerals as those in FIGS. 25 and 1 denote the same components. In FIG. 5A, reference numeral 50 denotes a semiconductor substrate constituting the sensor portion 1, and after the cover glass 31 is bonded by the adhesive layer 32, the film is thinned by a known technique. FIG. 2B shows a state in which the thinning process is completed, and reference numeral 51 denotes a thinned semiconductor substrate. In this image sensor form, in the process of reducing the thickness of the semiconductor substrate 50, the microlens group 30 functions as a stopper even when a mechanical force is applied from the lower side of the 50. The phenomenon that the main area of the image sensor is pushed into the cover glass side can be prevented. As a result, even when the thinning process is completed, the flatness of the back surface 51 is maintained as shown in FIG. 5B, and there is no inconvenience in the subsequent assembly process. Furthermore, it is possible to avoid the risk of the 51 being destroyed by a force applied suddenly from the back side of the 51 in the assembly process or the like. With this advantage, the semiconductor substrate 51 having a thickness of 100 μm or less can be easily realized, which can greatly contribute to the reduction in the height and size of the image sensor. In the figure, the reference numeral 31 is shown as a flat plate having a uniform thickness, but this is not restrictive. That is, only the region corresponding to the 31 microlens groups may be formed of a transparent glass material, the remaining region is a metal, and the glass material is hermetically sealed to the metal. This configuration is more suitable for applications that require high reliability. The space between the microlens group and the cover glass may be sealed with a vacuum. In such a configuration, after the adhesive is solidified, the microlens 31 and the microlens are more closely adhered to each other by the atmospheric pressure from the 31 side. In such a configuration, the thickness of the glass needs to be adjusted as appropriate so that necessary and sufficient deformation occurs corresponding to the pressure.

  FIG. 4 is a diagram showing a deformation state of the microlens, and the pixel region is enlarged and displayed. In FIG. 6A, 60 is a semiconductor substrate, 61 is a photodiode, 62 is a wiring layer that has a function of preventing light from being irradiated to a region other than the photodiode, 63 is a color filter, and 64 is a microlens. , 65 is a cover glass. FIG. 4A shows a state where the cover glass is in contact with the topmost part of the microlens. FIG. 5B shows a state where the cover glass crushes the topmost portion and forms a flat portion 66 on the microlens, and shows a state where the size of the flat portion is equal to the area of the photodiode. Yes. In such a state, the microlens curve shape portion 67 exists above the region other than the photodiode region. With this configuration, incident light is refracted in the curved shape, and the light is guided to the photodiode. FIG. 4C shows a state in which the microlens is deformed larger than FIG. In FIG. 5C, the flat portion 68 of the microlens is larger than the photodiode, and the curved shape 69 is smaller. In this configuration, light incident on the end of the flat portion is blocked by the wiring layer and cannot reach the photodiode, resulting in a decrease in photosensitivity. On the other hand, the curved shape 69 refracts incident light and guides the light to the photodiode. As a result, although the photosensitivity is reduced as compared with the configuration of FIG. 5B, the phenomenon that the main area of the image sensor is pushed into the cover glass in the thinning process of 60 can be prevented. The effectiveness of is maintained.

  FIG. 5 shows a process of manufacturing the image sensor according to FIG. In FIG. 2A, reference numeral 70 denotes a sensor portion manufactured by a known semiconductor process, and a microlens group 71 is provided on the surface. FIG. 2B shows a situation where a cover glass 73 provided with an adhesive layer 72 is arranged on the upper part of the 70. The mechanical position is controlled so that the 72 is positioned above the boundary region 74 of each image sensor indicated by the alternate long and short dash line, and then, as shown in FIG. Constructs a laminated structure. Such a laminated structure is placed in a high-temperature atmosphere using jigs and tools (not shown) that maintain the mutual positional relationship between 73 and 70, and the adhesive layer 72 is solidified. The initial thickness of the adhesive layer is set so that the lower surface of 73 comes into contact with the topmost part of the microlens by the solidification step, and further, the topmost part is pressed and deformed. As an example, the initial thickness is approximately the same as the thickness of the microlens (the distance from the top to the bottom of the microlens), but is not limited thereto.

  FIG. 6 shows another process example for manufacturing the image sensor of FIG. In the figure, the same reference numerals as those in FIG. 5 denote the same components. FIG. 2A shows a sensor portion 70 manufactured by a known semiconductor process and having a microlens group 71 disposed on the surface thereof. In FIG. 5B, an adhesive layer 82 is provided in the boundary region 74. The adhesive layer may be provided in a state of being patterned by a screen and a printing technique. Alternatively, the entire surface of 70 may be uniformly adhered by a technique such as spin coating, and the 82 having a desired pattern may be formed by a mask, an exposure technique, a selective removal technique, or the like. 83 in FIG. 8C is a cover glass, and unlike the case of FIG. 5 described above, the adhesive layer is not provided on the 83. After the 83 is arranged on the upper part of the 70, it is in close contact with the 70 to form a laminated structure as shown in FIG. Such a laminated structure is placed in a high-temperature atmosphere, and the adhesive layer 82 is solidified. The initial thickness of the adhesive layer is set so that the lower surface of 83 comes into contact with the topmost part of the microlens by the solidification step, and further, the topmost part is pressed and deformed. As an example, the initial thickness is approximately the same as the thickness of the microlens (the distance from the top to the bottom of the microlens), but is not limited thereto. The process example shown in FIG. 6 has an advantage that the relative positional relationship between the cover glass 83 and the sensor portion 70 does not need to be strictly controlled.

<Second image sensor configuration>
FIG. 7 is a diagram showing a second image sensor configuration according to the present invention. In the figure, the same reference numerals as those in FIG. 1 denote the same components. In the figure, reference numeral 91 denotes a cover glass, which is coupled to the sensor portion 1 by an adhesive layer 92. A depression 93 is provided on the surface of the 91 on the sensor portion side. Even in such a structure, due to the solidification of the 92, the depression 93 presses the top of the microlens group 30 to form a flat portion. This image sensor configuration is particularly effective when the thickness of the adhesive layer 92 after solidification is small and the deformation of the microlens is excessive. For the processing of the recess, a known method such as selective etching can be used. In addition, the depth of the depression is a condition that the topmost part of the microlens group can contact the surface of the 93 after the 92 is solidified.

<Third image sensor configuration>
FIG. 8 is a view showing a third image sensor form of the present invention, and the same reference numerals as those in FIG. 7 denote the same components. In the figure, a cover glass 101 is coupled to the sensor portion 1 via an adhesive layer 102. Reference numeral 103 denotes a dummy lens made of the same material and in the same process as the microlens group 30. In this image sensor configuration, the adhesive layer 102 is buried in the top 103 of the dummy lens 103. The planar shape of the dummy lens 103 is not necessarily the same as that of the microlens 30, and may have a base area larger than 30.

<Fourth image sensor configuration>
FIG. 9 is a view showing a fourth image sensor form of the present invention, and the same reference numerals as those in FIG. 8 denote the same components. In the figure, the surface of the cover glass 101 on the sensor portion 1 side has a convex portion 111, and the adhesive layer 102 is disposed and solidified without changing the shape of the dummy lens 103. In such a configuration, the topmost portion of the microlens group 30 is partially flattened by the presence of the convex portion 111.

<Fifth image sensor configuration>
FIG. 10 is a diagram showing a fifth image sensor form of the present invention, and the same reference numerals as those in FIG. 1 denote the same components. In this image sensor form, when processing to a thickness of approximately 100 μm or less in the thinning process of the semiconductor substrate 121, the central portion of the 121 is bent toward the governor glass, and the back surface of the 121 is flattened after the process is finished. Instead, the case where the convex part 122 occurs is illustrated. When the bending occurs, the topmost part of the microlens group 30 is strongly pressed against the cover glass 31 to increase the deformation amount, and the flat part 123 having a wider area is generated. However, when the thinning step is completed and the bending disappears, the flat portion 123 is separated from the lower surface of the cover glass 31, and a minute gap 124 is generated. Since the microlens group 30 serves as a stopper, the magnitude of the deflection is a much smaller value than the conventional example outlined in FIG. 24 and FIG. As a result, the height of the convex portion 122 on the back surface of the semiconductor substrate is small, and the flatness of the back surface is not maintained, but the influence can be reduced. Furthermore, even when sudden force is applied in the assembly process, the deformed microlens group 30 moves upward, and the movement stops on the back side surface of the cover glass 31, thereby preventing the 121 from being destroyed. it can. In such a configuration, the surface of the microlens is not necessarily in contact with the cover glass during the period in which the image sensor is operating, but the present invention such as reduction of connection failure, securing of connection reliability, and mechanical protection of the semiconductor substrate. Needless to say, this can contribute to the reduction in the height and size of the image sensor. In this image sensor configuration, the microlens group is not in contact with the cover glass before the thinning process of the substrate, and the contact and the microlens associated therewith during the thinning process. The case where the top portion is flattened and the contact state between the microlens group and the cover glass is eliminated after the thinning step is completed is also included.

  In FIG. 10, the material constituting the microlens has sufficient elastic characteristics, and even if the bending occurs in the thinning process 121, the bending amount can be absorbed, that is, the thinning process. It is possible to select the material so that the flat portion 123 on the surface of the microlens continues to come into contact with the back surface of the cover glass 31 when the deflection is eliminated after the end. Under such selection, the microlens group 30 functions as if it is a cushioning material, and the micro gap 124 shown in FIG. 10 cannot exist.

<Sixth image sensor configuration>
FIG. 11 shows a sixth image sensor form of the present invention. This figure is a conceptual view of a wafer state from above, in which a plurality of image sensors (two in the horizontal direction and two in the vertical direction are illustrated) are arranged. In the figure, 130 is a pixel, 131 is a microlens disposed on the pixel, 132 is a dot-and-dash line is the border of the image sensor in the vertical and horizontal directions, and 133 is for attaching a cover glass (not shown) to the top of the image sensor. It is an adhesive layer and is arranged in the vertical and horizontal directions along the boundary. In the drawing, 6 × 6 pixels are integrated in the sensor portion, and further, the entire surface of the image sensor region is covered with pixels. This figure corresponds to the plan view of FIG. 1 illustrated as a cross-sectional view, and is manufactured by the manufacturing method described above. However, in such a configuration, a sealed space is formed between the cover glass and the microlens group whose top is deformed flat. Since such a sealed space is normally filled with air at atmospheric pressure, the air expands due to the operating temperature of the image sensor. When the thermal expansion occurs, a force for pushing the cover glass upward from the bottom is generated, and deformation of the glass is induced. As a result, the contact state between the microlens and the cover glass in each pixel changes, a difference occurs in the light condensing action on the pixel, and an adverse effect on the image signal occurs. As an example, a gap is generated between the microlens and the cover glass at the periphery of the sensor portion, and a larger gap is generated at the center of the sensor portion.

<Seventh image sensor configuration>
FIG. 12 shows a seventh image sensor form of the present invention for eliminating the effect of such thermal expansion. In the figure, the same reference numerals as those in FIG. 11 denote the same components. In the figure, reference numeral 140 denotes an adhesive layer similar to 133, but there are regions 141 and 142 in which no adhesive layer is present. In the regions 141 and 142, the cover glass and the image sensor are not coupled to each other, the above-described sealed space is not configured, and the space is opened. As a result, since the air can freely enter and exit through the regions 141 and 142, the influence of the thermal expansion of the air can be avoided even if the operating temperature of the image sensor changes. The size (length) of the regions 141 and 142 can be selected as appropriate, but it is desirable that the regions 141 and 142 be small so that foreign matters such as dust and contaminants do not enter the space between the image sensor and the cover glass through the regions.

<Eighth image sensor configuration>
FIG. 13 is a diagram showing an eighth image sensor form of the present invention, and is a conceptual sectional view of the image sensor. In the figure, the same reference numerals as those in FIG. 1 denote the same components. In this image sensor configuration, the adhesive layer 32 does not have the regions 141 and 142 having no adhesive layer as illustrated in FIG. In the figure, reference numeral 151 denotes a second through hole for opening a space 155 between the sensor portion 1 and the cover glass 31. The through hole is created in the same process as the first through hole 150 filled with the conductive material. It is desirable that the positions of the 141 holes are appropriately selected in the periphery of the image sensor and do not affect the operation of the image sensor. After the through holes 150 and 151 are formed, the conductive material is filled, but by setting the size of the hole 151 larger than 150, the inside of 151 is not completely filled, and the side wall portion Only the conductive material is covered. After the filling process is completed, the opening 152 is formed from the surface side of the sensor portion. The formation process is easily realized by a combination of a photoresist, an exposure technique, selective etching, and the like. Other procedures can be used to form 152 and 151. For example, the procedure is such that after the opening 152 is provided, the cover glass 31 is bonded to the top of the 1, the semiconductor substrate 2 is thinned, the through holes 150 and 151 are formed, and then the conductive material is filled. Can be mentioned. In the configuration of FIG. 13, since the space 155 between the sensor portion 1 and the cover glass 31 is opened to the external space by the 152 and 151, it results from the sealed space generated in the form of the image sensor illustrated in FIG. The effect can be eliminated. Further, the image sensor configuration of FIG. 13 has an advantage that foreign matters such as dust and contaminants are less likely to enter the space 155 between the image sensor and the cover glass as compared to the image sensor configuration of FIG. The advantage is that, after the image sensor is mounted, the 155 is open to the surface on which the 10 is provided, so that foreign matter and contaminants are unlikely to reach the lower opening of the 151. ing.

<Ninth image sensor configuration>
FIG. 14 shows a ninth image sensor form of the present invention. In FIG. 11, the same reference numerals as those in FIG. 11 denote the same components. The image sensor form of the figure is similar to the image sensor form of FIGS. 8 and 9 described above, and is characterized in that an adhesive layer is provided on the first dummy lens. In FIG. 14, reference numeral 161 denotes a first dummy lens group arranged in the boundary 132 region, which is a structure that is created in the same process as the microlens arranged in the pixel region and has a similar shape to the microlens. . In the figure, reference numeral 162 denotes an adhesive layer on the dummy lens. In this image sensor form, the arrangement pitch of the dummy lens group is substantially the same as the pixel arrangement pitch, and the size is also substantially the same as the pixel area. In such a configuration, since there is a gap between the individual dummy lenses, there is an advantage that the effect due to the sealed space described with reference to FIG. 11 can be eliminated.

  The structure referred to as the first dummy lens in this image sensor form is disposed at a position unrelated to the photodiode of the pixel, and therefore does not have an optical function such as condensing, and is exclusively used. Only the shape is used. As described above, it is created in the same process as the microlens, and the geometric shape of the dummy lens is almost the same as that of the microlens. Therefore, in this specification, the name of the dummy lens is given for convenience.

<Tenth image sensor configuration>
FIG. 15 shows a tenth image sensor configuration according to the present invention. The same reference numerals as those in FIG. 11 denote the same components. In the figure, reference numeral 171 denotes a pixel region in which the 130 and 131 are arranged, and reference numeral 172 denotes a peripheral circuit region having no photoelectric conversion function. In an image sensor, for example, a signal processing circuit that amplifies a signal from a photodiode or converts it into a time-series signal, a drive circuit that sequentially selects photodiodes, a power supply circuit, various control signals supplied from the outside such as a digital camera A logic circuit for processing is integrated on the same semiconductor substrate. The peripheral circuit is disposed in the region 172 and is usually covered with a light shielding layer in order to avoid malfunctioning with incident light. When the present invention is applied to an image sensor having such a configuration, the mechanical contact between the cover glass and the microlens group is only a part of the entire region (usually referred to as a chip or a die) divided by 132. That is, it is achieved only in the pixel region indicated by 171. As a result, the contraction force generated when the adhesive layer indicated by 133 is solidified presses the cover glass to the chip side over the entire surface of the chip, but since there is no microlens in the area 172, It is assumed that the chip is deformed more greatly toward the chip side. In the image sensor configuration shown in FIG. 15, the second dummy lens group 173 is also arranged in the peripheral circuit region 172 in order to prevent such deformation bias. The 173 is created in the same process as the microlens in the photosensitive area. Like the first dummy lens, the second dummy lens does not have an optical function and only uses the shape. In addition, the arrangement pitch of the second dummy lenses 173 is shown to be approximately equal to the pixel pitch, but is not limited thereto. Furthermore, the planar size of the dummy lens may be arbitrarily set regardless of the size of the pixel.

<11th image sensor configuration>
FIG. 16 shows an eleventh image sensor form of the present invention, and the same reference numerals as those in FIG. 15 denote the same components. In FIG. 14, similarly to FIG. 14, the first dummy lens group 181 is arranged in the boundary region of each image sensor, and the adhesive layer 182 is provided on the lens. This image sensor configuration has both the advantage of eliminating the effect due to the sealed space outlined in FIG. 14 and the advantage of eliminating the bias of cover glass deformation outlined in FIG.

<Twelfth image sensor configuration>
FIG. 17 is a diagram showing a twelfth image sensor form according to the present invention. The same reference numerals as those in FIG. 16 denote the same components. In this image sensor form, the arrangement pitch of the second dummy lenses arranged in the peripheral circuit region 172 is different from the pixel pitch, and further, the size of the second dummy lens is the size of the micro lens in the pixel. An example in which different dummy lenses are included is shown. As described above, since the purpose of the second dummy lens is to prevent the deformation of the cover glass, the effect does not change sensitively with respect to the arrangement pitch and size. The selection of the pitch and the size should be appropriately determined from the arrangement state of electronic elements such as transistors in the peripheral circuit region.

<13th image sensor configuration>
FIG. 18 is a view showing a thirteenth image sensor form according to the present invention. The same reference numerals as those in FIG. 11 denote the same components. In the figure, reference numeral 201 denotes a third dummy lens group disposed around the pixel group, which is disposed between the pixel and the boundary region of the image sensor. The reference numeral 201 is a structure having a planar spread created in the same process as the microlens in the photosensitive region. The third dummy lens, like the first dummy lens and the second dummy lens described above, does not have an optical function and only uses the shape. Such a configuration is an example, and other arrangement methods, for example, a configuration in which the third dummy lens group is arranged only in the horizontal direction or the vertical direction, a configuration in which the arrangement pitch of the third dummy lenses is different from the pixel pitch, A configuration including a configuration in which the size of the third dummy lens is different from the size of the microlens in the pixel is also included in this image sensor configuration. This image sensor form is intended to ensure the uniformity of contact between the cover glass and the microlens by disposing a third dummy lens around the pixel. For this purpose, the arrangement and size of the third dummy lens are not limited, and various arrangements and sizes can be selected as appropriate.

  FIG. 19 is a view showing a process for manufacturing the structure of the image sensor form shown in FIG. 18, and the same reference numerals as those in FIG. 18 denote the same components. In FIG. 5A, a sensor portion 211 provided with a microlens 131 and a third dummy lens 201 is prepared, while a cover glass 212 provided with an adhesive layer 133 is also prepared. The 211 and 212 are stacked after the mechanical position is adjusted so that the boundary 213 coincides, as shown in FIG. In the state shown in FIG. 5B, the adhesive layer 133 is solidified in a high temperature atmosphere while maintaining the positional relationship of the components. In the solidification process, the adhesive layer softens and flows to the periphery of the third dummy lens, but the area or thickness of the adhesive layer must be adjusted so that the adhesive does not diffuse to the pixel region where the microlens 131 is present. I must. Such a situation is illustrated in FIG. Through this process, the topmost parts of the dummy lens 201 and the microlens 131 are brought into contact with the cover glass, and further, the topmost part is flattened by being pressed by a mechanical force generated with the shrinkage and solidification of the adhesive layer. The

<14th image sensor configuration>
FIG. 20 is a diagram showing a fourteenth image sensor form of the present invention. For convenience, only the corner portions of the four image sensors are shown in FIG. 18A, and the same numbers as those in FIG. 18 indicate the same components. In the figure, third dummy lenses 221 and 222 are arranged around the pixel region and between the boundary 132. As shown in the figure, in the image sensor form, the planar size of the 221 is set to be larger than the planar size of the 222. Such a difference in size will be described in detail below with reference to FIGS. FIGS. 7B and 7C are structural cross-sectional views showing manufacturing steps of the microlens and the third dummy lens. The micro lens and the dummy lens are manufactured using the same type of resin. More specifically, after the semiconductor process is completed, a resin serving as a material for the lens is formed on the entire surface of the semiconductor substrate with a uniform thickness using a technique such as spin coating. Subsequently, patterns 223, 224 and 225 similar to the shape of the lens are formed by the exposure technique [(b) of FIG. In the figure, 223 is a third dummy lens having a large area (221 in FIG. 11A), 224 is a third dummy lens having a small area (222 in FIG. 22A), and 225 is a microlens. [131] corresponds to [131 in FIG. The thicknesses of 223 and 224 are indicated by 226 and 227, respectively, but they are equal because they are immediately after coating and patterning. Next, such a structure is solidified in a high temperature atmosphere. Although the solidification process is different from that of the resin to be used, it is generally solidified by evaporation of a solvent contained in the original resin and a chemical reaction. Since the viscosity decreases and the fluidity increases at the initial stage of the solidification, the surface has a generally arc shape due to the action of surface tension. FIG. 2C is a structural cross-sectional view at the stage where the solidification is completed. In the figure, the dummy lens and the microlens having a small area have a circular arc shape as shown by 229 and 230, respectively, while the dummy lens having a large area has a shape of the lens as shown by 228. Only the periphery is arcuate, leaving a flat portion in the center. This difference in shape results in 232 becoming larger, as indicated by 232 and 231 as the thickness (height) of each dummy lens. The semiconductor substrate, that is, the sensor portion that has undergone such a manufacturing process is laminated in the processes shown in FIGS. As shown in FIG. 4D, a cover glass 235 on which an adhesive layer 234 is selectively disposed is opposed to a sensor portion 233 provided with third dummy lenses 228 and 229, and a boundary 236 is formed. The relative positional relationship is adjusted with reference to. The 233 and the 235 are laminated, and the adhesive layer 234 is solidified in a high temperature atmosphere. In this solidification step, pressure is applied to the 233 and the 235 from above and below, but by appropriately setting the pressure, the lower surface of the 235 is kept in contact with the upper surface of the 228. Is possible. As outlined in (b) and (c) of the figure, the height of the 229 is slightly larger than the height of the 228, so that the top of the 229 is crushed in the contact state of the 228 and 235. As a result, a flat portion is generated. FIG. 4E is a structural sectional view showing such a state. In this image sensor form, by providing a difference in the area between the microlens and the third dummy lens, a minute height difference is generated after the lens group is solidified, and the height difference is used to cover the lens group. It is characterized by carrying out laminated lamination of glass. In particular, in the method of flattening the topmost part of the microlens, it is important to set the distance between the cover glass and the sensor part. If the distance is too small, the generally circular arc shape of the microlens disappears, while the distance If is excessively large, the flat portion of the microlens is not generated. As described above, this image sensor configuration has a great advantage that the appropriate interval can be easily set by setting the area of the lens group.

<15th image sensor configuration>
FIG. 21 is a diagram showing a fifteenth image sensor configuration according to the present invention. The same reference numerals as those in FIG. 20 denote the same components. In FIG. 21A of this image sensor form, a weir 240 is provided around the pixel region and the boundary 132 instead of the third dummy lens of FIG. The width of the weir is set larger than the size of the microlens. The weir is created in the same process as the microlens 131 manufacturing process. This manufacturing process is shown in FIGS. (B) and (c). In FIG. 4B, reference numerals 243 and 244 denote patterns formed on the sensor portion by spin coating, exposure, selective etching, or the like, using the material constituting the microlens. The 243 corresponds to the weir 240, and the 244 corresponds to the microlens 131. Since 243 and 244 are created in the same process, the thicknesses (heights) indicated by 245 and 246 are equal. The figure which processed the said structure in high temperature atmosphere, and dried and solidified the said material is shown by the figure (c). In FIG. 8C, 247 and 248 are shapes after solidification corresponding to the 243 and 244, respectively. Since the size of the 248 is smaller than the width of the 247, the surface of the 248 is a generally arcuate curved surface. On the other hand, the surface of the 247 has a flat portion at the center portion, and the shoulder portion is a curved surface having a substantially arc shape. Since the geometric planes of the 248 and the 247 are different, the thicknesses (heights) indicated by 250 and 249 are not equal, and the 250 is larger. By utilizing such a difference in thickness, the top portion of the microlens 131 is in contact with the lower surface of the cover glass and a flat portion is formed, as outlined in FIGS. 20 (d) and 20 (e). The width of the weir 240 is required to be larger than the microlens, and the position is not particularly limited. In addition, as suggested in FIG. 21 (a), the vertical and horizontal widths of the weirs do not necessarily have to coincide with each other as long as the width of both is larger than the size of the microlens. Furthermore, there may be a configuration in which only one of the 240 vertical and horizontal directions is formed.

<16th image sensor configuration>
FIG. 22 is a diagram showing a sixteenth image sensor form of the present invention. In FIG. 4A, 260 is a sensor portion, 261 is a microlens in the pixel region, and 262 is the above-described third dummy lens or weir. Reference numeral 264 denotes an opening provided at the boundary 263. The opening is formed by selectively removing an oxide film or a resin film that forms the surface of the 260. For reference, the opening 264 is provided to facilitate scribing when a chip is cut from a wafer in a general integrated circuit manufacturing process. In this image sensor form, it is not always necessary to ensure such ease. A cover glass 267 partially provided with an adhesive layer 265 is disposed facing the 260, and the 265 is adjusted so that the planar position of the 265 coincides with the opening 264. After the adjustment, the 260 and the 267 are laminated, and the solidifying process of the 265 is performed. FIG. 2B is a sectional view of the structure after the solidification process is completed. By the solidification step, the lower surface of the cover glass 267 comes into contact with the 262, and at the same time, the topmost portion of the 261 is crushed to form a flat portion. In this image sensor configuration, since the opening 264 functions as a storage reservoir for the adhesive layer 265, there is an advantage that the possibility that excessive adhesive flows out to the pixel region is reduced. Although the thickness and width of the adhesive layer 265 are appropriately selected, it is necessary to prevent the excessive adhesive from flowing to the pixel region in the state shown in FIG.

  The relationship between the adhesive layer and the dummy lens group in the above-described plural image sensor embodiments of the present invention will be described below. In the plurality of image sensor configurations, the presence of the first dummy lens, the second dummy lens, the third dummy lens, and the weir is extremely important. However, there is a difference that an adhesive layer exists on the first dummy lens, but no adhesive layer exists on the second dummy lens, the third dummy lens, and the weir. In the component in which the adhesive layer is not present, the adhesive layer is present around the component, and the component is in direct contact with the lower surface of the governor glass in the solidifying process of the adhesive layer. That is, the component functions as a stopper that defines the vertical position of the cover glass during the solidification process. As a result, there is an advantage that variation does not easily occur in flattening the topmost portion of the condensing microlens existing in the pixel region. Note that a configuration in which an adhesive layer is disposed on the component can also be adopted if the variation is allowed and a characteristic difference between pixels is allowed.

  In this specification, a configuration in which an image sensor is covered with a cover glass is described. If the term “cover glass” in the present specification is described more strictly, it will be “a flat plate having at least the property of being transparent in a specific light wavelength region”. That is, the cover glass may have not only a function of covering the image sensor to protect the image sensor but also various functions. For example, when the image sensor is applied to a digital camera, an infrared cutoff filter function that optically removes near-infrared light that affects the imaging effect can be given. In the case of an image sensor for black-and-white imaging, the near-infrared light energy contained in the incident light may be actively used to increase the apparent light sensitivity. In such a case, the cover glass needs to be transparent to a wavelength range from visible light to near infrared light. In addition, in the case where the image sensor is applied to an infrared imaging device intended for thermal image detection, an infrared filter that removes visible light that affects the imaging effect can be used. In these application examples, the filter is formed on the front surface or the back surface of the cover glass. As the structure of the filter, there are a method of directly forming an interference filter in which multiple layers of metal and dielectric thin films are laminated on the glass surface, and a method of stretching a plastic film laminated in multiple stages and stretching it on the glass surface. Further, as an example of a function added to the cover glass, there is a mounting of a microlens. In this case, a microlens is formed on the front or back surface of the cover glass so as to be equal to the pixel pitch. In such a configuration, there may be a case where the above-described microlens of the sensor portion is not necessary. In some cases, two lens systems may be formed. Another example of the function added to the cover glass is a color filter. In this case, forming a color filter array equal to the pixel pitch on the front or back surface of the cover glass. In such a configuration, the color filter array may be directly formed on the glass surface by a well-known method, or the color filter array may be formed on a glass thin plate different from the glass and then attached to the cover glass. good. The term “cover glass” in this specification includes these functional examples exemplified in this section, and both are included in the present invention.

  In this specification, as shown in FIG. 1, a configuration of an image sensor considered as a standard configuration is described. However, there are various configurations of the image sensor, and the present invention can be applied to all of them as long as a condensing microlens is mounted in the pixel region. For example, the micro lens in the pixel area is not transparent and has a color filter function, the configuration of a black and white image sensor without a color filter, and the space between the micro lens and the cover glass is vacuum. So-called back-illuminated image sensor configuration for increasing sensitivity, so-called amplifying image in which each pixel has an amplifying function or the photodiode itself has an amplifying function for increasing sensitivity The structure of the sensor, a CCD image sensor that reads signals from the pixels via horizontal and vertical registers, and the image sensor itself has a stacked structure, and each layer has an imaging function, signal processing function, memory function, There are three-dimensional configurations where output control functions are assigned. A plurality of such configuration examples are all included in the present invention.

<17th image sensor configuration>
FIG. 23 is a diagram showing a seventeenth image sensor form of the present invention. This figure shows a state where the image sensor after the assembly process is assembled as a camera. In the figure, 270 is an image sensor, 271 is an imaging lens, 272 is a focal position adjusting mechanism of the 271, 273 is a housing of the lens, and 274 is a substrate on which a camera-related electronic member is mounted. Reference numerals 275 and 276 denote adhesive layers for fixing the camera housing 273 to the substrate 274. The adhesive layer is flexible and has a function of pressing the image sensor 270 against the substrate 274. As another configuration example, one of 275 and 276 may be an adhesive layer, and the remaining may be a flexible layer. This image sensor form is characterized in that the 275 and 276 also have a function of pressing the cover glass constituting the 270 against the top of the microlens. With this configuration, the contact state between the cover glass and the microlens is stabilized, and the condensing state of incident light in each pixel can be made uniform over the entire pixel region.

  The present invention is not limited to an image sensor, and can be applied to an optical device such as a display that requires a process for thinning a base substrate.

It is a figure which shows the cross-section of a semiconductor image sensor. (First image sensor configuration) It is a figure which shows the cross-sectional shape of a micro lens. It is a figure which shows the thin film formation process of the image sensor of FIG. It is a figure which shows the deformation | transformation condition of a micro lens. It is a figure which shows the manufacturing process of the image sensor of FIG. It is a figure which shows the other manufacturing process of the image sensor of FIG. It is a figure which shows the cross-section of an image sensor which has a recessed part in a cover glass. (Second image sensor configuration) It is a figure which shows the cross-section of an image sensor with a dummy lens. (Third image sensor configuration) It is a figure which shows the cross-section of the image sensor which has a convex part in a cover glass. (Fourth image sensor configuration) It is a figure which shows the cross-section of an image sensor in which a micro gap exists between cover glasses. (Fifth image sensor configuration) It is a figure which shows the planar structure of an image sensor. (Sixth image sensor configuration) It is a figure which shows the planar structure of the image sensor without an adhesive layer in a part of boundary. (Seventh image sensor configuration) It is a figure which shows the cross-sectional structure of an image sensor with the through-hole of an air vent. (Eighth image sensor configuration) It is a figure which shows the planar structure of the image sensor which has a dummy lens in a boundary. (9th image sensor form) It is a figure which shows the planar structure of the image sensor which has a dummy lens in a peripheral circuit area | region. (10th image sensor form) It is a figure which shows the planar structure of the image sensor which has a dummy lens in a boundary and a peripheral circuit area | region. (Eleventh image sensor form) It is a figure which shows the planar structure of the image sensor which has multiple types of dummy lenses in a peripheral circuit area | region. (Twelfth image sensor configuration) It is a figure which shows the planar structure of the image sensor which has a dummy lens around a chip | tip. (13th image sensor form) It is a figure which shows the manufacturing process of the image sensor of FIG. It is structural sectional drawing which shows a planar structure of an image sensor with a dummy lens from which a magnitude | size differs, and a part of manufacturing process. (14th image sensor form) It is structural sectional drawing which shows the planar structure of the image sensor which has a dam around a chip | tip, and a part of its manufacturing process. (15th image sensor form) It is a sectional view showing a part of the manufacturing process of the image sensor. (16th image sensor form) It is a figure which shows the cross-section when an image sensor is assembled as a camera. (17th image sensor form) It is a figure which shows the cross-section of the conventional semiconductor image sensor. It is a figure which shows the thin film formation process of the conventional semiconductor image sensor.

Explanation of symbols

1, 70, 211, 233, 260 Sensor portion 2, 50, 60, 121 Semiconductor substrate 3, 61 Photodiode 4, 63 Color filter 5, 30, 64, 71, 131, 230, 261 Microlens 6, 31, 65 73, 83, 91, 101, 212, 235, 267 Cover glass 7, 32, 72, 82, 92, 102, 133, 140, 162, 182, 234,
265, 275, 276 Adhesive layer 8 Pad region 9 Conductive material 10 Connection terminal group 20, 74, 132, 213, 236, 263 Boundary 21, 51 Semiconductor substrate with small thickness 40 Arc-shaped curves 41, 66, 68, 123 Flat part of the center part of the micro lens 42, 67, 69 Arc-shaped part of the peripheral part of the micro lens 62 Wiring layer 93 Depression 103 Dummy lens 111, 122 Convex part 124 Gap 130 Pixel 141, 142 Area without adhesive layer 150, 151 Through hole 152, 264 Aperture 155 Space 161, 181 First dummy lens 171 Pixel area 172 Peripheral circuit area 173 Second dummy lens 201, 221, 222, 228, 229 Third dummy lens 223, 224, 225, 243, 244 Pattern 226, 227, 231, 23 Shape 262 after the thickness 240 weir 247 and 248 solidifies 245,246,249,250 third dummy lens or weir 270 image sensor 271 imaging lens 272 focus position adjusting mechanism 273 lens housing 274 substrate

Claims (3)

  In a semiconductor image sensor in which a flat plate having at least a characteristic that is transparent in a specific light wavelength region is provided on the surface, and each pixel is equipped with a microlens, the microlens and the lower surface of the flat plate A semiconductor image sensor comprising a microlens partly flattened by contact with the surface.   The semiconductor image sensor according to claim 1, wherein a plurality of structures having a shape similar to the microlens or a structure having a planar extension is arranged in a region other than the photosensitive region where the pixels are arranged, A semiconductor image sensor, wherein the structure is in contact with a lower surface of the flat plate. 3. The semiconductor image sensor according to claim 1, wherein a plurality of types of the structures having different planar dimensions are arranged.