JP2004319610A - Image sensor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve a problem that colors are mixed in an image sensor performing photoelectric conversion of a primary color light having a wavelength selected by one kind of color filter, and a problem that slanting light or misregistration occurs in an image sensor where spectral characteristics of a desired primary color light are enhanced by more than one kind of filter. <P>SOLUTION: Photoelectric conversion areas 12B, 12G and 12R of pixels for respective primary color lights are formed in a silicon substrate 11 at such depths as corresponding to the light absorption characteristics of silicon in the wavelength region of an incident primary color light by utilizing the light absorption characteristics of the silicon substrate 11 such that the spectral characteristics for the incident primary color light become identical substantially. Wavelength selectivity of a light incident to the pixel of each primary color light can thereby be enhanced and since one kind of color filter 14B, 14G or 14R is employed for each primary color light, problems of slanting light and misregistration can be solved. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an image sensor, and more particularly, to a color image sensor using a photodiode.
[0002]
[Prior art]
When a color image sensor is manufactured using a semiconductor such as a CCD (charge transfer device) or a CMOS imager, each primary color light of R (red), G (green), and B (blue) is provided for each pixel. For example, as shown in FIG. 4, the transmission filters are arranged in a so-called checkerboard pattern in which different primary color light transmission filters are adjacent to each other, and the filters are printed and arranged using an exposure mask or the like. .
[0003]
However, the spectral characteristic of such a filter for transmitting primary color light is a characteristic in which the wavelength region where a spectral transmittance equal to or higher than a predetermined value is obtained is considerably widened, as described in Patent Document 1, for example. The wavelength regions in which spectral transmittances equal to or higher than a predetermined value are obtained overlap considerably between the G filter and the G transmission filter and between the G transmission filter and the R transmission filter. As a result, color mixing in which each color mixes occurs, causing deterioration of the color signal.
[0004]
Therefore, in order to avoid such color mixing, as in the invention described in Patent Document 1, in FIG. 5, the first filter whose spectral characteristic is indicated by a dotted line I, and the spectral characteristic is indicated by a one-dot chain line II. A method of designing a combined filter having a steep peak at a specific wavelength as shown by a solid line III by combining the second filter and the second filter is conceivable.
[0005]
[Patent Document 1]
JP-A-9-127408 (FIGS. 34 and 37)
[0006]
[Problems to be solved by the invention]
In the conventional image sensor described in Patent Document 1, when a color image sensor is configured, as shown in FIG. 6, a B pixel 2B, a G pixel 2G, and an R pixel formed on a silicon substrate 1 are formed. A first B filter 5B, a first G filter 5G, and a first R filter 5R are provided above each of the 2Rs via a wiring layer and an interlayer film 3, and an interlayer film 4 is further provided thereon. The second filter 7B for B, the second filter 7G for G, and the second filter 7R for R are provided through the protective film 6.
[0007]
Here, the first B filter 5B and the second B filter 7B are arranged in the vertical direction with respect to the B pixel 2B and have two types of spectral characteristics different from each other. Then, for the normal light incident on the vertical path indicated by the dotted line IV, a combined filter characteristic having a steep peak with respect to the wavelength of the blue light described above is given.
[0008]
Similarly, the first G filter 5G and the second G filter 7G are arranged in the vertical direction with respect to the G pixel 2G and have two types of spectral characteristics different from each other. Then, when they are combined, they have a synthetic filter characteristic having a steep peak with respect to the wavelength of green light, and the first R filter 5R and the second R filter 7R are provided with an R pixel. It is arranged in the vertical direction with respect to 2R, has two kinds of different spectral characteristics from each other, and has a combined filter characteristic having a sharp peak with respect to the wavelength of red light when they are combined.
[0009]
However, in the conventional image sensor having the structure shown in FIG. 6, when light is obliquely incident on the B pixel 2B as shown by a dashed line V in FIG. Since the light passes through the filter 7G for B, a filter characteristic synthesized so as to have a steep peak with respect to the wavelength of blue light cannot be obtained, and light in a useless wavelength range enters the B pixel 2B. However, noise occurs in the B signal obtained by the B pixel 2B, and the S / N deteriorates. The same applies to the case where light is obliquely incident on the other G pixels 2G and R pixels 2R.
[0010]
Further, in the conventional image sensor, if there is a positional shift between two types of filters or between a filter and a pixel, there is light that passes through only one of the filters, and the spectral characteristics are deteriorated, and the primary color signal obtained by imaging is deteriorated. Such troubles occur.
[0011]
The present invention has been made in view of the above points, and has as its object to provide an image sensor having higher spectral selectivity and higher spectral characteristics with less color mixing based on a principle different from the conventional one.
[0012]
[Means for Solving the Problems]
In order to achieve the above object, the present invention provides a first primary color light pixel for photoelectrically converting a first primary color light wavelength-selected by a first color filter from incident light, and a second color filter from incident light. A second primary color light pixel that photoelectrically converts the second primary color light whose wavelength has been selected according to the above, and a third primary color light that photoelectrically converts the third primary color light whose wavelength has been selected by the third color filter from the incident light. In an image sensor in which a plurality of pixels are arranged in a two-dimensional matrix or a one-dimensional straight line, each of a first primary color light pixel, a second primary color light pixel, and a third primary color light pixel is The semiconductor device is formed on the silicon substrate at a depth corresponding to the light absorption characteristic of silicon in the wavelength region of the incident primary color light so that the spectral characteristics of the incident primary color light are substantially the same.
[0013]
The present invention utilizes the fact that the light absorption characteristic of silicon has a predetermined light absorption characteristic in accordance with each wavelength region of the three primary color lights, and each of the pixels for the primary color light is converted into a spectral characteristic with respect to the incident primary color light. Are formed on the silicon substrate at a depth corresponding to the light absorption characteristics of silicon in the wavelength region of the incident primary color light so that the wavelengths are substantially the same as each other. In addition, even light incident obliquely on the surface of the silicon substrate can be converted into light transmitted through one color filter.
[0014]
BEST MODE FOR CARRYING OUT THE INVENTION
Next, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a structural sectional view of an embodiment of an image sensor according to the present invention. In the figure, on a silicon substrate 11, B photoelectric conversion regions 12B, G photoelectric conversion regions 12G, and R photoelectric conversion regions 12R are provided at different positions in the horizontal direction (parallel to the substrate surface). Are formed at different depths from each other.
[0015]
Here, the photoelectric conversion region 12B for B is formed at a depth D1 from the surface of the silicon substrate 11 and the photoelectric conversion region 12G for G is formed at a depth D1 to (D1 + D2) from the surface of the silicon substrate 11 for a reason to be described later. Are formed within the range. The R photoelectric conversion region 12R is formed in a range from the depth (D1 + D2) to (D1 + D2 + D3) from the surface of the silicon substrate 11.
[0016]
A wiring layer and an interlayer film 13 are formed on the silicon substrate 11 on which the B photoelectric conversion region 12B, the G photoelectric conversion region 12G, and the R photoelectric conversion region 12R are formed. On the interlayer film 13, the B filter 14B, the G filter 14G, and the R filter 14R are respectively perpendicular to the corresponding B photoelectric conversion region 12B, G photoelectric conversion region 12G, and R photoelectric conversion region 12R. They are arranged and formed so as to be aligned. The B filter 14B is a filter that selects blue light from the incident light, the G filter 14G is a filter that selects green light from the incident light, and the R filter 14R is a filter that selects red light from the incident light. These are filters for selecting wavelengths, which are covered and protected by a common light-transmitting protective film 15.
[0017]
In this embodiment, the three primary color light filters, that is, the B filter 14B, the G filter 14G, and the R filter 14R are used together with the three primary color light filters to utilize the light absorption characteristics of the silicon substrate 11. The feature is that the spectral characteristics of the filter are improved.
[0018]
That is, it is already known from the literature (Hiroo Yonezu, “Optical Communication Device Optics”, Engineering Books, p. 323) that the light absorption characteristics of silicon are the characteristics indicated by Si in FIG. From the light absorption characteristics Si of silicon in FIG. 2 in which the vertical axis indicates the absorption coefficient and the horizontal axis indicates the wavelength, it can be understood how much incident light enters the silicon substrate for each wavelength. Thus, in the present embodiment, the photodiode for each primary color light has wavelength selectivity by utilizing the light absorption characteristics of silicon.
[0019]
Here, since the characteristics of the photodiode (photoelectric conversion region) of the image sensor are usually the same for the R, G, and B pixels, in FIG. The absorption coefficient in the blue light wavelength region at the center is the largest, the absorption coefficient in the red light wavelength region around the wavelength of 0.64 μm is the smallest, and the absorption coefficient in the green light wavelength region around the wavelength of 0.53 μm. 1, the photoelectric conversion region 12B for blue light having the largest absorption coefficient is formed on the silicon substrate 11 as shown in FIG. Is formed at the deepest, and the G photoelectric conversion region 12G is formed at an intermediate depth.
[0020]
The value of the depth depends on where the peak wavelengths of R, G, and B are taken. In FIG. 1, for example, D1 is 0 to 0.5 μm, D2 is 0.5 to 1.2 μm, and D3 is 1.2 to 3 μm. It is set such as 0.0 μm.
[0021]
Thus, according to the present embodiment, the B photoelectric conversion region 12B has a sharp peak with respect to the wavelength of blue light due to the combination with the B filter 14B and the light absorption characteristics of the silicon substrate 11. Blue light to which filter characteristics have been imparted is incident and photoelectrically converted. Similarly, the G photoelectric conversion region 12G is provided with a green light having a filter characteristic having a steep peak with respect to the wavelength of the green light by a combination of the G filter 14G and the light absorption characteristics of the silicon substrate 11. Is incident and photoelectrically converted, and the R photoelectric conversion region 12R has a filter characteristic having a steep peak with respect to the wavelength of red light due to a combination with the R filter 14R and the light absorption characteristics of the silicon substrate 11. The applied red light is incident and photoelectrically converted.
[0022]
Therefore, according to the present embodiment, the wavelength selectivity of the light incident on each primary color light pixel is improved, so that the influence of color mixing is higher than that of a conventional image sensor using one type of primary color light selection filter. Can be reduced. Further, as shown in FIG. 1, in the present embodiment, the primary color light selection filter has only one layer, so that obliquely incident light passes through a plurality of different primary color light selection filters and enters the photoelectric conversion region. Therefore, the problem of obliquely incident light can be solved, and the problem of displacement does not occur.
[0023]
Next, a method for manufacturing the image sensor according to the present embodiment will be described with reference to element cross-sectional views in each step of FIG. In the figure, the same components as those in FIG. 1 are denoted by the same reference numerals. Usually, an N-type photoelectric conversion region is often provided on a P-type substrate in a CCD, a CMOS imager, or the like. Therefore, in order to realize the image sensor of FIG. 1, an N-type impurity and a P-type impurity are implanted as shown in FIG. 3 to adjust the N-type depth. The configuration shown in FIG. 3 is a configuration often used in a CCD or CMOS imager in which charges generated by photoelectric conversion in an N-type layer are moved to another diffusion layer by an N-MOSFET.
[0024]
First, as shown in FIG. 3A, a gate oxide film 24 is coated on the entire surface of the P-type silicon substrate 11 and, for example, polycrystalline polysilicon is formed on the entire surface thereof, and then the gate electrodes 25 ₁ and 25 ₂ are formed. , except with the polycrystalline silicon to be 25 ₃ oxide film 24 thereunder, polycrystalline polysilicon and oxide film 24 is removed by etching. Thereafter, respectively an N-type diffusion layer 21 ₃ of the N-type diffusion layer 21 _2, the depth of the N-type diffusion layer 21 _1, the depth of the depth D1 on the P-type silicon substrate 11 (D1 + D2) (D1 + D2 + D3) by ion implantation and, also, the same depth of the N type diffusion layer (or CCD transfer _path) 22 _1, 22 to form a 2, 22 _3. Here, the N-type diffusion layers 21 ₁ , 21 ₂ , and 21 ₃ having different depths can be easily formed at a desired depth by a known method by adjusting the implantation energy and the implantation amount, the thermal diffusion time, the thermal diffusion temperature, and the like. Can be formed.
[0025]
Subsequently, as shown in FIG. 3 (B), a P-type impurity having a higher concentration than the silicon substrate 11 is ion-implanted into the surfaces of the N-type diffusion layers 21 ₂ and 21 ₃ to form P ⁺ diffusion regions 26 ₁ and 26 _2. To form Here, the implantation energy and implantation amount of P-type impurity, the thermal diffusion time, by adjusting the thermal diffusion temperature, ^{P +} diffusion region ₂₆ in a known manner 1, 26 ₂ each depth D1, in (D1 + D2) Can be formed.
[0026]
Then, as shown in FIG. 3 (C), B filter to cover the wiring layer and the interlayer film 13 on a silicon substrate 11, corresponding to the N-type diffusion layer ₂₁ 1 of the _above, 21 2, 21 ₃ position The filter 14B, the filter 14G for G, and the filter 14R for R are formed by a known method, and finally covered with the protective film 15, thereby completing the manufacture of the image sensor. Here, the N-type diffusion layers 21 ₁ , 21 ₂ , and 21 ₃ correspond to the B photoelectric conversion region 12B, the G photoelectric conversion region 12G, and the R photoelectric conversion region 12R shown in FIG. 11 and a photodiode. In this way, it is possible to manufacture the image sensor according to the present embodiment in which the pixels of each primary color have more wavelength selectivity and can reduce the influence of color mixing.
[0027]
It should be noted that the present invention provides an image sensor having a configuration in which three types of pixels for separately performing photoelectric conversion of three primary color lights are arranged in a two-dimensional matrix as shown in FIG. Of course, it can be applied to any of the image sensors.
[0028]
【The invention's effect】
As described above, according to the present invention, the light absorption characteristics of silicon in the wavelength region of incident primary color light are set so that each of the pixels for primary color light has substantially the same spectral characteristic with respect to the incident primary color light. By forming on the silicon substrate at a depth corresponding to the wavelength, the wavelength selectivity of light incident on each primary color light pixel is improved, so that the influence of color mixing can be reduced, and the primary color light incident on each primary color light pixel can be reduced. Since there is only one type of color filter for selecting the wavelength, a primary color signal due to oblique incident light or a positional shift of the color filter as in a conventional image sensor that obtains a desired primary color light by a combined spectral characteristic of two or more types of color filters. The problem such as deterioration of S / N can be prevented.
[Brief description of the drawings]
FIG. 1 is a structural sectional view of an embodiment of the present invention.
FIG. 2 is an explanatory diagram of light absorption characteristics of silicon.
FIG. 3 is a cross-sectional view of an element in each step for explaining a manufacturing method according to an embodiment of the present invention.
FIG. 4 is a diagram illustrating the arrangement of filters for transmitting primary color light of the image sensor.
FIG. 5 is a spectral characteristic diagram of each of two filters and a spectral characteristic diagram obtained by combining them;
FIG. 6 is a structural sectional view and a problem explanatory view of a conventional image sensor.
[Explanation of symbols]
11 P-type silicon substrate 12B B photoelectric conversion region 12G G photoelectric conversion region 12R R photoelectric conversion region 13 Wiring layer and interlayer film 14B B filter 14G G filter 14R R filter 15 Protective films 21 ₁ , 21 ₂ , 21 ₃ N-type diffusion layer _{22 1,} 22 2, 22 ₃ diffusion layer or the CCD transfer paths _{25 1,} 25 2, 25 ₃ gate electrode

Claims (1)

A first primary color light pixel that photoelectrically converts the first primary color light wavelength-selected by the first color filter from the incident light, and a second primary color light wavelength-selected by the second color filter from the incident light A second primary color light pixel for photoelectrically converting and a third primary color light pixel for photoelectrically converting the third primary color light whose wavelength has been selected by the third color filter from the incident light are two-dimensional matrix or primary. In image sensors arranged in a plurality of original linear shapes,
Each of the first primary color light pixel, the second primary color light pixel, and the third primary color light pixel has incident primary color light such that spectral characteristics with respect to the incident primary color light are substantially the same. An image sensor formed on a silicon substrate at a depth corresponding to the light absorption characteristics of silicon in the wavelength region described above.

Images