JP6911353B2 - Manufacturing method of solid-state image sensor - Google Patents

Description

The present invention relates to a solid-state image sensor having a microlens for each pixel and a method for manufacturing the same.

In recent years, image pickup devices have become widely used with the expansion of image recording, communication, and broadcasting contents. Various types of image pickup devices have been proposed, but image pickup devices incorporating solid-state image sensors, which are compact, lightweight, and have come to be stably manufactured, are used as digital cameras and digital videos. It is becoming popular.

The solid-state image sensor has a plurality of photoelectric conversion elements that receive an optical image from an object to be photographed and convert the incident light into an electric signal. The types of photoelectric conversion elements are roughly classified into CCD (charge-coupled device) type and CMOS (complementary metal oxide semiconductor) type. In addition, there are two types of photoelectric conversion elements, a linear sensor (line sensor) in which the photoelectric conversion elements are arranged in a row and an area sensor (plane sensor) in which the photoelectric conversion elements are two-dimensionally arranged vertically and horizontally. It is roughly divided into. In any of the sensors, the larger the number of photoelectric conversion elements (number of pixels), the more precise the captured image. Therefore, in recent years, a method for inexpensively manufacturing a solid-state image sensor having a large number of pixels has been studied.

In addition, by providing a color filter function that transmits light of a specific wavelength in the path of light incident on the photoelectric conversion element, color solid-state imaging as a single-plate color sensor that makes it possible to obtain color information of an object. Elements are also widespread. The color solid-state image sensor can collect color-separated image information by forming a pattern of one pixel by a specific colored transparent pixel corresponding to one photoelectric conversion element and arranging a large number of them regularly. The colors of the colored transparent pixels are three primary colors consisting of three colors of red (R), green (G), and blue (B), or cyan (C), magenta (M), and yellow (Y). Complementary color systems consisting of are generally used, and in particular, three primary color systems are often used.

One of the important issues in the performance required for a solid-state image sensor is to improve the sensitivity to incident light. In order to increase the amount of information in an image taken by a miniaturized solid-state image sensor, it is necessary to make the photoelectric conversion element, which is a light receiving unit, finer and highly integrated. However, when the photoelectric conversion elements are highly integrated, the area of each photoelectric conversion element becomes smaller and the area ratio that can be used as the light receiving part also decreases, so that the amount of light that can be taken into the light receiving part of the photoelectric conversion element becomes smaller, which is effective. Sensitivity is reduced.

As a means for preventing such a decrease in the sensitivity of the miniaturized solid-state image sensor, in order to efficiently capture the light into the light receiving portion of the photoelectric conversion element, the light incident from the object is collected for each pixel. A technique has been proposed in which a microlens that leads to a light receiving portion of a photoelectric conversion element is formed on the photoelectric conversion element in a uniform shape. By condensing light with a microlens and guiding it to the light receiving portion of the photoelectric conversion element, it is possible to increase the apparent aperture ratio of the light receiving portion and improve the sensitivity of the solid-state image sensor.

Here, as a method for forming a microlens, there are a flow lens type and a dry etching transfer type. In the flow lens type, first, a lens forming layer made of a transparent and heat-flowing acrylic photosensitive resin, which is a material of a microlens, is formed, and the lens forming layer is divided into a plurality of pixel units (rectangular pattern). The pattern is formed by the photolithography method. Next, the lens forming layer is heated to form a microlens in each pixel unit.

In the dry etching transfer type, first, a lens mother is made of an acrylic transparent resin, which is a material for a microlens, and a resist material having alkali solubility, photosensitivity, and heat flow property is used on a lens forming layer having a flat upper surface. Form a mold layer. Next, the lens master is formed by performing a photolithography step and a heat flow step on the lens master layer. That is, this matrix layer forming step is performed by the same method as the flow lens type microlens forming method. Next, the lens cambium is dry-etched using the lens matrix as a mask to transfer the shape of the lens cambium to the lens cambium to form a microlens.
Further, in recent years, various methods for manufacturing a microlens using a grayscale mask have been proposed. For example, Patent Document 1 proposes a method for designing a grayscale mask and a method for manufacturing a grayscale mask and a microlens, which can obtain a stable microlens shape even if the exposure amount fluctuates to some extent.

Japanese Unexamined Patent Publication No. 2014-174456

In recent years, the number of pixels in solid-state image sensors has increased, and high-definition solid-state image sensors with more than several million pixels have been required. The deterioration of image quality due to the increase in noise has become a problem.
In order to improve the sensitivity of the solid-state image sensor, it is also required to reduce crosstalk. Crosstalk is a phenomenon in which light that should originally be incident on a certain color is incident on the adjacent color due to the influence of the difference in the refractive index of the pigments of each color. Due to the influence of crosstalk, a color having a low refractive index loses light due to an adjacent color having a high refractive index, so that the amount of light to the light receiving portion is reduced and the sensitivity is lowered.
That is, the reduction of crosstalk improves the focusing efficiency of the microlens.
An object of the present invention is to provide a solid-state image sensor having a microlens whose focusing efficiency is improved by reducing crosstalk, and a method for manufacturing the same.

The solid-state image sensor according to the first aspect of the present invention includes a semiconductor substrate, a plurality of photoelectric conversion elements formed on the semiconductor substrate, and photoelectric conversion elements arranged in a matrix in the plane of the semiconductor substrate. A flattening layer formed on a plurality of photoelectric conversion elements and a plurality of color filters formed on the flattening layer, which are arranged in the same matrix as the plurality of photoelectric conversion elements and emit light in each wavelength band. A plurality of color filters for transmitting light and a plurality of microlenses formed on the plurality of color filters are provided.
The edges of the plurality of microlenses are connected to each other in a valley shape, and the cross-sectional shape of the valley portion forming the connecting portion perpendicular to the semiconductor substrate surface is the first cross section along the row of the matrix and the column of the matrix. It is V-shaped in the second cross section along and in the third cross section at 45 degrees to the rows and columns.

A method for manufacturing a solid-state image sensor, which is the second aspect of the present invention, is performed on a plurality of photoelectric conversion elements formed on a semiconductor substrate and arranged in a matrix in the plane of the semiconductor substrate via a flattening layer. It includes a color filter forming step of forming each of the above color filters, and a microlens forming step of forming a plurality of microlenses on a plurality of color filters after the color filter forming step.
The microlens forming step includes a master layer forming step, a lens master forming step, and a heat flow step. The master layer forming step is a step of forming a lens master layer made of a transparent resin having photosensitive and heat flow properties on a plurality of color filters. The lens matrix forming step is a step of forming a plurality of lens masters in the lens master layer by a photolithography method using a gray tone mask, and a plurality of gaps are provided between the edges of adjacent lens masters. This is a process of making the shape of the lens matrix of No. 1 into a shape in which the height from the color filter is higher and the spread on the color filter surface is smaller than that of a plurality of microlenses. The heat flow step is a step of heating a plurality of lens master molds to form a plurality of microlenses.

According to the present invention, there is provided a solid-state image sensor having a microlens with reduced crosstalk and improved light collection efficiency, and a method for manufacturing the same.

It is a top view explaining the solid-state image sensor of one Embodiment of this invention. It is a figure which shows the solid-state image sensor of one Embodiment of this invention, and is the figure corresponding to the first cross section (the II cross section of FIG. 1) along the row of the matrix of a color filter. It is a figure which shows the solid-state image sensor of one Embodiment of this invention, and is the figure corresponding to the second cross section (II-II cross section of FIG. 1) along the row of the matrix of a color filter. It is an enlarged view of the IV part of FIG. It is a figure explaining the operation of this invention, and is the figure corresponding to the 3rd section (the section III-III of FIG. 1) which becomes 45 degrees with respect to the row and column of the matrix of a color filter. It is a figure explaining the method of one Embodiment of this invention.

Hereinafter, embodiments of the present invention will be described, but the present invention is not limited to the embodiments shown below. In the embodiments shown below, technically preferable limitations are made for carrying out the present invention, but this limitation is not an essential requirement of the present invention.

<Structure>
As shown in FIGS. 1 to 3, the solid-state image sensor 6 of the present embodiment has a photoelectric conversion element 2, a flattening layer 3, a plurality of color filters 4, and a plurality of microlenses 5 on a semiconductor substrate 1. It is formed by stacking in this order. In addition, in FIG. 1, in order to make it easy to understand the arrangement of the plurality of photoelectric conversion elements 2 and the plurality of color filters 4, other configurations in the solid-state image pickup element 6 are omitted.
The semiconductor substrate 1 is a substrate for mounting the photoelectric conversion element 2. The photoelectric conversion element 2 converts the light incident through the microlens 5 and the color filter 4 into electric charges. The flattening layer 3 flattens the upper surface of the semiconductor substrate 1, which is the mounting surface of the microlens 5.
The plurality of color filters 4 are each formed on the plurality of photoelectric conversion elements 2 via the flattening layer 3. The plurality of color filters 4 have a role of transmitting light having a specific wavelength in the path of light incident on the photoelectric conversion element 2. In the present embodiment, the plurality of color filters 4 transmit any one of the three colors of red (R), green (G), and blue (B), and the three colors are Bayer-arranged. Is.

The plurality of microlenses 5 are formed on the plurality of color filters 4, respectively. The plurality of microlenses 5 are made of a transparent resin, and the material thereof is usually a resin such as an acrylic resin, and transparent is preferable.
Further, the edges of the plurality of microlenses 5 are connected to each other in a valley shape next to each other. Further, the cross-sectional shapes of the valleys forming the connecting portions 5a of the plurality of microlenses 5 perpendicular to the semiconductor substrate surface are the first cross-section along the rows of the matrix, the second cross-section along the rows of the matrix, and the rows and columns. It is V-shaped in the third cross section which is 45 degrees with respect to the relative.
Further, the line showing the surface 5b excluding the connecting portion 5a of the plurality of microlenses in the first cross section, the second cross section, and the third cross section is a parabola as shown in FIGS. 2 and 3, but is an arc or a sine. It may be a waveform.

FIG. 4 shows an enlarged view of a valley portion (hereinafter, also referred to as “valley”) forming the connecting portion 5a of the plurality of microlenses 5. When the valley is arcuate, the radius of curvature R at the lowest point L of the valley is expressed by the following equation. In the equation, f (x) is a function indicating the shape curve of the valley portion forming the connecting portion 5a, and a is the x coordinate at the lowest point L of the valley portion.

In the first cross section, the second cross section, and the third cross section of the plurality of microlenses 5, for example, as shown in FIG. 5A, when the radius of curvature R of the valley between the plurality of microlenses 5 is large, the color is transparent. The light incident on the microlens in the vicinity of the blue of the pixel is changed to the colored transparent pixel of green having a refractive index larger than that of blue in the color filter layer. As a result, there is concern that the effects of crosstalk will increase.
On the other hand, for example, as shown in FIG. 5B, when the radius of curvature R of the valley between the plurality of microlenses 5 is small, a large amount of light incident on the adjacent microlenses on the blue color of the colored transparent pixels is large. Light enters the photoelectric conversion element 2 through the color filter 4 and the flattening layer 3 without changing the optical path. As a result, the influence of crosstalk is reduced and the light collection efficiency can be improved.
An example shown in FIG. 5B, in which the radius of curvature R is 50 nm or less, is included in "the cross-sectional shape of the valley portion perpendicular to the semiconductor substrate surface is V-shaped". Further, as a matter of course, an example in which the lowest point of the valley is a contact point between straight lines (in the above equation, when the lowest point L is, for example, f (x) = | x | x = 0). Is also included in "the cross-sectional shape perpendicular to the semiconductor substrate surface of the valley is V-shaped".

<Manufacturing method>
Next, a method of manufacturing the solid-state image sensor 6 of the present embodiment will be described with reference to FIG.
First, a flattening layer 3 (not shown in FIG. 6) and a layer of a color filter 4 are sequentially laminated on a semiconductor substrate 1 (not shown in FIG. 6) on which a photoelectric conversion element 2 is formed on a surface portion (color). Filter forming process). In the color filter forming step, three types of color filters corresponding to any of RGB are arranged in a matrix (for example, in a predetermined pattern as shown in FIG. 1) and laminated on a plurality of photoelectric conversion elements.

After the color filter forming step, a plurality of microlenses 5 are formed on the plurality of color filters 4 (microlens forming step). In the microlens forming step, first, as shown in FIG. 6A, a transparent resin having photosensitivity and heat flow is applied on a plurality of color filters 4 to a predetermined thickness to form a lens matrix layer. 10 is formed (matrix layer forming step).
Next, as shown in FIG. 6B, the lens matrix layer 10 is exposed to the lens matrix layer 10 using the gray tone mask 11 based on the photolithography method, and then developed and baked to obtain a plurality of lenses. The master mold (lens master mold) 12 of the microlens 5 of the above is formed (lens master mold forming step). In this step, a gap is provided between the edges of the adjacent lens masters 12, and the shapes of the plurality of lens masters 12 are higher than those of the plurality of microlenses 5 from the color filter 4 and are formed on the color filter surface. The shape should be small in spread.

Further, in this step, the lens master mold is used by using a gray tone mask capable of arbitrarily changing the mask transmittance gradation according to the shape of the lens master mold 12 according to the desired shape of the microlens 5. It becomes easy to control the shape of the twelve. The shading of the mask transmittance gradation is achieved by a partial difference in density per unit area of dots having a small diameter that cannot be resolved by the light used for exposure.
In the lens master mold 12, the gap between the edges of the adjacent lens master molds 12 is set to 50 nm or more and 250 nm or less in consideration of the amount of heat flow in the heat flow step which is the next step.

Next, the master mold 12 of the microlens is melted by heat. That is, the microlens 5 is formed by heat-flowing the master molds 12 of the plurality of microlenses (heat flow step). By forming the microlens 5 by the heat flow, it is possible to form the microlens 5 in which the valley between the adjacent microlenses is narrowed.

<Effect of this embodiment>
According to the solid-state image sensor of the present embodiment, in the microlens forming step, by forming the matrix of the microlens by the photolithography method using a gray tone mask, the shape of the microlens can be easily controlled, and each individual image pickup is performed. By making it possible to select and form the optimum microlens shape for each element, there is an effect that the light collection efficiency to the photoelectric conversion element is enhanced.
Further, by forming the microlens by the heat flow, it is possible to narrow the valley between the adjacent microlenses, so that there is an effect that the light collection efficiency to the photoelectric conversion element can be improved.

Hereinafter, the first embodiment will be described.
As the semiconductor substrate, a silicon wafer having a thickness of 0.75 mm and a diameter of 20 cm was used. A photoelectric conversion element was formed on the upper surface of the silicon wafer, and a flattening layer was formed on the uppermost layer by spin coating using a thermosetting type acrylic resin coating liquid.
Next, three color resists of red (R), green (G), and blue (B) were used on the flattening layer, and three color filter layers were sequentially formed by a photolithography method. The film thickness of each color filter layer was 0.5 to 0.8 μm. In the pixel arrangement of the color filter layer, a green (G) filter is provided every other pixel, and a red (R) filter and a blue (B) filter are provided every other row between the green (G) filters, so-called. It was a Bayer arrangement.

Next, an alkali-soluble and photosensitive acrylic transparent resin was applied onto the color filter layer at a film thickness of 1.0 μm, and heated at 90 ° C. for 2 minutes to harden the film.
Then, a matrix of microlenses was formed on the acrylic transparent resin by a photolithography method using a gray tone mask. As the gray tone mask of Example 1, a photomask designed so that the shape of the microlens becomes a parabolic shape after the heat flow was used.
Next, in the baking process, the mother mold of the microlens formed by the photolithography method using a gray tone mask was heat-flowed. The baking conditions at this time were the treatment in a three-step step of 160 ° C. → 180 ° C. → 250 ° C.

When the shape of the microlenses formed in the first embodiment was measured with a scanning probe microscope, the edges of the plurality of microlenses were connected to each other in a valley shape and perpendicular to the semiconductor substrate in the valley portion forming the connecting portion. The cross-sectional shape is V-shaped in both the cross-section (first and second cross-sections) and 45-degree cross-section (third cross-section), and the line indicating the surface of the part excluding the connecting part is a parabolic line. I confirmed that there was.
Further, when the light receiving efficiency of the solid-state image sensor formed in Example 1 and the conventional product was measured, it was confirmed that the solid-state image sensor formed in Example 1 had a better result of about 5.3%. did. The conventional product compared in Example 1 is a microlens formed directly by a photolithography method using a gray tone mask without passing through a master mold, with a design having the same shape as that in Example 1. In this conventional product, the edges of a plurality of microlenses are connected to each other in a valley shape, and the line indicating the surface of the portion excluding the connecting portion is a parabola, and the cross section of the valley portion forming the connecting portion is perpendicular to the semiconductor substrate. The shape was arcuate, and the radius of curvature was 194 nm in the cross-sectional direction and 104 nm in the 45-degree cross-sectional direction.

Hereinafter, the second embodiment will be described.
As the semiconductor substrate, a silicon wafer having a thickness of 0.75 mm and a diameter of 20 cm was used. A photoelectric conversion element was formed on the upper surface of the silicon wafer, and a flattening layer was formed on the uppermost layer by spin coating using a thermosetting type acrylic resin coating liquid.
Next, three color resists of red (R), green (G), and blue (B) were used on the flattening layer, and three color filter layers were sequentially formed by a photolithography method. The film thickness of each color filter layer was formed to be 0.5 to 0.8 μm. In the pixel arrangement of the color filter layer, a green (G) filter is provided every other pixel, and a red (R) filter and a blue (B) filter are provided every other row between the green (G) filters, so-called. It was a Bayer arrangement.

Next, an alkali-soluble and photosensitive acrylic transparent resin was applied onto the color filter layer at a film thickness of 1.0 μm, and heated at 90 ° C. for 2 minutes to harden the film.
Then, a matrix of microlenses was formed on the acrylic transparent resin by a photolithography method using a gray tone mask. As the gray tone mask of the second embodiment, a photomask designed so that the shape of the microlens becomes an arc shape after the heat flow is used.
Next, in the baking process, the mother mold of the microlens formed by the photolithography method using a gray tone mask was heat-flowed. The baking conditions at this time were the treatment in a three-step step of 160 ° C. → 180 ° C. → 250 ° C.

When the shape of the microlens formed in Example 2 was measured with a scanning probe microscope, the edges of the plurality of microlenses were connected to each other in a valley shape and perpendicular to the semiconductor substrate in the valley portion forming the connecting portion. The cross-sectional shape is V-shaped in both the cross section (first cross section and second cross section) and the 45 degree cross section (third cross section), and the part excluding the connecting part is formed in a spherical shape. It was confirmed.
Further, when the light receiving efficiency of the solid-state image sensor formed in Example 2 and the conventional product was measured, it was confirmed that the solid-state image sensor formed in Example 2 had a better result of about 4.8%. did. In the conventional product compared in Example 2, the microlens was directly formed by the photolithography method using a gray tone mask without passing through the master mold, with the same shape design as in Example 2. In this conventional product, the edges of a plurality of microlenses are connected to each other in a valley shape, the portion excluding the connecting portion is spherical, and the cross-sectional shape of the valley portion forming the connecting portion perpendicular to the semiconductor substrate is arcuate. The radius of curvature was 212 nm in the cross-sectional direction and 114 nm in the 45-degree cross-sectional direction.

Hereinafter, the third embodiment will be described.
As the semiconductor substrate, a silicon wafer having a thickness of 0.75 mm and a diameter of 20 cm was used. A photoelectric conversion element was formed on the upper surface of the silicon wafer, and a flattening layer was formed on the uppermost layer by spin coating using a thermosetting type acrylic resin coating liquid.
Next, three color resists of red (R), green (G), and blue (B) were used on the flattening layer, and three color filter layers were sequentially formed by a photolithography method. The film thickness of each color filter layer was formed to be 0.5 to 0.8 μm. In the pixel arrangement of the color filter layer, a green (G) filter is provided every other pixel, and a red (R) filter and a blue (B) filter are provided every other row between the green (G) filters, so-called. It was a Bayer arrangement.

Next, an alkali-soluble and photosensitive acrylic transparent resin was applied onto the color filter layer at a film thickness of 1.0 μm, and heated at 90 ° C. for 2 minutes to harden the film.
Then, a matrix of microlenses was formed on the acrylic transparent resin by a photolithography method using a gray tone mask. As the gray tone mask of Example 2, a photomask designed so that the shape of the microlens becomes a sinusoidal shape after the heat flow was used.

Next, in the baking process, the mother mold of the microlens formed by the photolithography method using a gray tone mask was heat-flowed. The baking conditions at this time were the treatment in a three-step step of 160 ° C. → 180 ° C. → 250 ° C.
When the shape of the microlens formed in Example 3 was measured with a scanning probe microscope, the edges of the plurality of microlenses were connected to each other in a valley shape and perpendicular to the semiconductor substrate in the valley portion forming the connecting portion. The cross-sectional shape is V-shaped in both the cross-section (first and second cross-sections) and 45-degree cross-section (third cross-section), and the line showing the surface of the part excluding the connecting part is a sinusoidal waveform. I confirmed that.

Further, when the light receiving efficiency of the solid-state image sensor formed in Example 3 and the conventional product was measured, it was confirmed that the solid-state image sensor formed in Example 3 had a better result of about 3.7%. did. In the conventional product compared in Example 3, a microlens was directly formed by a photolithography method using a gray tone mask without passing through a master mold, with a design having the same shape as in Example 3. In this conventional product, the edges of a plurality of microlenses are connected to each other in a valley shape, and the line showing the surface of the portion excluding the connecting portion has a sinusoidal waveform and is perpendicular to the semiconductor substrate of the valley portion forming the connecting portion. The cross-sectional shape was arcuate, and the radius of curvature was 243 nm in the cross-sectional direction and 120 nm in the 45-degree cross-sectional direction.
The results of Examples 1 to 3 are summarized in Table 1 below.

1 Semiconductor substrate 2 Photoelectric conversion element 3 Flattening layer 4 Color filter 5 Microlens 5a Microlens connection 5b Surface excluding microlens connection 6 Solid-state image sensor 10 Lens master layer 11 Gray tone mask 12 Lens master R Radius of curvature of the valley between adjacent microlenses L The lowest point of the valley between adjacent microlenses

Claims (2)

A color filter forming step of forming a plurality of color filters on a plurality of photoelectric conversion elements formed on the semiconductor substrate and arranged in a matrix in the plane of the semiconductor substrate via a flattening layer, and the color. After the filter forming step, the microlens forming step of forming a plurality of microlenses on the plurality of color filters is included.
The microlens forming step is
A master layer forming step of forming a lens master layer made of a transparent resin having photosensitivity and heat flow on the plurality of color filters.
A step of forming a plurality of lens masters on the lens master layer by a photolithography method using a gray tone mask, wherein a gap is provided between the edges of the adjacent lens masters, and the plurality of lens masters are formed. A lens matrix forming step in which the shape of the lens is higher than that of the plurality of microlenses from the color filter and the spread on the color filter surface is small.
A heat flow step of heating the plurality of lens masters to form the plurality of microlenses, and
Have,
The gray tone mask has the shape of the lens matrix.
After the heat flow step, the shape of the microlens
The edges of the plurality of microlenses are connected to each other in a valley shape, and the cross-sectional shape of the valley portion forming the connecting portion perpendicular to the semiconductor substrate surface is the first cross section along the row of the matrix, that of the matrix. A photomask design has been made to control the second cross section along the column and the third cross section at 45 degrees to the row and the column so as to have a V shape.
A method for manufacturing a solid-state image sensor, which does not perform an exposure step on the edges of the plurality of microlenses between the lens matrix forming step and the heat flow step. The method for manufacturing a solid-state image sensor according to claim 1 , wherein the gap is 50 nm or more and 250 nm or less.

Images