JPS6242426B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (1) Field of Application of the Invention The present invention relates to a solid-state imaging device used in a television camera or the like.

(2) An overview of the conventional technology and its problems are shown in Figures 1 and 2.
This will be explained using the diagram and FIG.

FIG. 1 shows an example of the configuration of a typical two-dimensional solid-state imaging device. A pixel array including a photodiode 1 and a MOS transistor 2 is sequentially scanned by a horizontal scanning circuit 9 and a vertical scanning circuit 10 each comprising a MOS shift register, for example, by sequentially conducting the MOS transistors 3 and 2, respectively. The electric charge generated by the light accumulated in the photodiode 1 is extracted from the output terminal 8 through the signal line 6 and the signal line 7, and the image signal received by the pixel is extracted as an electric signal. There may be a plurality of signal lines 7 and output terminals 8 depending on the purpose.

FIG. 2 shows the cross-sectional structure of a typical pixel. For convenience of explanation, an N-channel type imaging device using electrons as signal charges will be described below, but the following explanation can also be applied to a P-channel type device by simply reversing the conductivity type and polarity.

A photodiode is formed by a Si substrate 11 made of a P-type Si single crystal and an n-type diffusion layer 12, and at the same time, the n-type diffusion layer 12 serves as a source with a gate electrode 13 made of polycrystalline Si, for example, and a gate electrode 13 under the gate electrode 13. A MOS field effect transistor is formed together with the thinned SiO ₂ film 16 and the n-type diffusion layer 14 as a drain. In order to lower the electrical resistance of the n-type diffusion layer 14, an electrode 17 made of a metal such as Al is usually provided and used as the signal line 6 in FIG. Further, the SiO ₂ 16 film is usually made thick to suppress the generation of unnecessary parasitic capacitance outside the pixel.

When the light 15 is incident, the n-type diffusion layer 12 and Si
Electron-hole pairs are generated in the substrate 11, and the electrons flow into the n-type diffusion layer 12 as signal charges, forming n-hole pairs.
pn junction capacitance 1 between type diffusion layer 12 and Si substrate 11
Accumulated by 8. When a positive scanning pulse is applied to the gate electrode 13, the electrons are attracted to the n-type diffusion layer 14 (drain) which is at a positive potential, and the electrons are extracted to the n-type diffusion layer 14, as explained with reference to FIG. It is led to the output end 8.

As a result, the potential of the n-type diffusion layer 12 becomes a positive potential, and the light 15 is turned off until the next positive scanning pulse is applied.
The electrons generated by this process continue to accumulate, and the potential decreases.

These conventional solid-state imaging devices, especially the photodiodes that form their core, have the following four structural disadvantages, which prevent them from being put into practical use, despite the strong demand for solid-state imaging devices. .

First, in order to improve resolution, it is necessary to make the pixels smaller, which inevitably reduces the area of the n-type diffusion layer 12, resulting in a smaller junction capacitance 18, which reduces the amount of electrons that can be stored. In other words, the signal charge is reduced. In particular, the area of the n-type diffusion layer 14 from which signal charges are taken out is made as small as possible regardless of the size of the pixel, and due to this wrinkle, the n-type diffusion layer 12 is further reduced significantly, and the signal charge in FIG. The number of n-type diffusion layers 14 connected to the line 6 increases as the number of pixels increases,
Each n-type diffusion layer 14 has a junction capacitance 19, and due to the large capacitance added to the parasitic capacitance of the signal line 6 and the signal line 7, the electrical signal appearing at the output terminal 8 becomes extremely small and is buried in electrical noise, making detection difficult. It becomes impossible. Therefore, with the current technology, it is not possible to obtain a resolution comparable to that of television images.

This problem can be solved by improving the junction capacitance 18 by about seven orders of magnitude. Junction capacitance 18 increases as the impurity concentration of Si substrate 11 increases. However, at the same time, the junction capacitance 19 also increases, which cannot be solved by this method.

The next problem concerns spectral properties. The absorption of light by Si differs depending on the wavelength. Three primary colors red (wavelength 0.65
Figure 3 shows the absorption characteristics of green (0.55 μm) and blue (0.45 μm) as R, G, and B, respectively. The blue light B creates electron-hole pairs near the surface of Si, and the red light R penetrates deep into the Si and creates electron-hole pairs. On the surface of Si, there is a high probability that recombination occurs and electron-hole pairs disappear, which makes them insensitive.As a result, photodiodes using Si have high sensitivity to red light (long wavelength light). Furthermore, the number of electron-hole pairs is inversely proportional to the wavelength for the same light energy, and as a result, longer wavelength light becomes more advantageous, and the spectral characteristics of the photodiode become distorted with higher sensitivity on the long wavelength side. Therefore, when electrical signals obtained from a conventional image sensor are reproduced into an image,
The blue parts of the original picture become darkened, and the red parts become whitish, creating an unnatural appearance.

The third problem relates to the case of strong incident light. In FIG. 2, when the light 15 is strong, the junction capacitance 18 is naturally saturated, and as shown in FIG. 3, many electron-hole pairs are generated even in the deep part of the Si substrate 11. Among these electrons, which are minority carriers, do not necessarily go toward the n-type diffusion layer, but also diffuse in the structural direction and are injected into the adjacent n-type diffusion layer. As a result, the optical signal spreads between many pixels, and in the reproduced image, it becomes a large white bright area that collapses the screen (this is called a blooming phenomenon). This phenomenon is noticeable in conventional devices.

A fourth problem relates to color imaging. In conventional devices, it is not possible to change the spectral characteristics of the photodiode. Therefore, three primary color filters are layered on each of the three solid-state imaging devices.
It is extremely difficult to align the images of each color, and the optical system that distributes the images to each imaging device is expensive and large, which detracts from the great advantage of miniaturizing the camera as a solid-state imaging device.

Naturally, it is desirable to be able to obtain a color signal with a single imaging device, and for this reason it is conceivable to place a color filter on the n-type diffusion layer 12, but materials used in semiconductor devices that can be processed with high precision are However, at present, it is difficult to form color filters on the n-type diffusion layer 12 with high precision, partly because the production of color filters itself is difficult.

Furthermore, since pixels for each color are required to obtain color signals, the number of pixels inevitably increases. As a result, for the same reason as the first problem, it is difficult to obtain color signals using conventional solid-state imaging devices.

(3) Purpose of the invention The purpose of the present invention is to dramatically increase the signal charge capacity of the pixels of a solid-state imaging device, correct the spectral characteristics,
It suppresses dispersion (blooming) of optical signals incident on other pixels, makes it possible to realize a high-resolution solid-state imaging device, and also provides means for realizing a single-chip color imaging device that obtains color signals with a single imaging device. It is something to do. The present invention will be explained below with reference to Examples.

Note that the present invention is intended only to improve pixels, and the following description will be made only with respect to pixels.

(4) Examples Hereinafter, the present invention will be explained in detail with reference to examples.

FIG. 4 shows one embodiment of the invention. This pixel differs from the conventional pixel shown in FIG. 2 in that a P ⁺ buried layer 25 with a high concentration of P type impurities is provided below the n type diffusion layer 22. The P ⁺ buried layer 25 is the gate electrode 23
Similar effects can be obtained on the side surfaces of the n-diffused layer 22, except for the area below. The same applies to the following examples.

A potential diagram for atoms along line A-A' in FIG. 4 is shown by curve 36 in FIG. In the figure, 32 is an n-type diffusion layer 22, 35 is a P ⁺ buried layer 25,
31 is the Si substrate 21, 33 is the n-type diffusion layer 22 and P ⁺
This corresponds to a depletion layer formed between the buried layers 25.
Region 35 of P ⁺ buried layer 25 is region 3 of Si substrate 21
1, the electrons generated inside the Si substrate 21 are driven into the Si substrate 21 as shown by the arrow 37, and the electrons stored in the n-type diffusion layer 22 are generated between the surface and the highest point 34. limited to electrons that have been

As a result, the problem of electron inflow (blooming) caused by strong light incident on other pixels is solved.

It can be seen from FIG. 3 that the position of the highest point 34 determines the spectral characteristics of the photodiode. If we quantify it, if the depth from the surface of the highest point 34 is x, then
The number N of electrons generated by the incident light is given by the following equation for the light intensity I: N=I·η·(1−e ⁻ 〓 ^x ) (1). η is the probability of generation of an electron-hole pair with light of each wavelength, and is proportional to the wavelength. α is the light absorption coefficient. α=0.34, 0.71, and 2.41 μm ⁻¹ for red light (wavelength 0.65 μm), green light (wavelength 0.55 μm), and blue light (wavelength 0.45 μm), respectively.

The position of the highest point 34 roughly corresponds to the point with the highest impurity concentration in the P ⁺ buried layer 25, and when the concentration is uniform, it is roughly at the center. From Figure 3 and equation (1), x
It can be seen that the smaller the value, the relatively higher the short wavelength light sensitivity becomes.

If there is no recombination of electron-hole pairs on the surface, the number of electrons obtained per unit energy by light of each wavelength will be approximately equal at around x=3 μm. The rate of recombination varies depending on processing technology, but
If the value is as high as can be obtained with the current technology, x
= 1.5 to 2 μm, the wavelength dependence of spectral sensitivity almost disappears.

The junction capacitance is the n-type diffused layer 22 and the P ⁺ buried layer 25
It is determined by the impurity concentration of the P ⁺ buried layer 25 near the boundary. The higher the concentration, the higher the junction capacitance. However, the electrical withstand voltage of the junction decreases. Practically speaking, an appropriate impurity concentration is about 10 ¹⁶ to 10 ¹⁸ cm ^-3 , preferably about 1×10 ¹⁷ cm ^-3 . 1×10 ¹⁷ cm ^-3 , the Si substrate 21 contains Si with an impurity concentration of about 10 ¹⁴ cm ^-3.
If a wafer is used, this P ⁺ buried layer 2
The presence of 5 improves the junction capacitance by about 30 times.
Electrical withstand voltage is 15V or higher. n-type diffusion layer 24 and Si
There is no change in the junction capacitance with the substrate 21.

As can be seen from the above description, all three problems in monochrome (black and white) solid-state imaging devices have been solved by the present invention. The following description relates to a single-chip color imaging device.

FIG. 6 shows another embodiment of the invention. Only photodiodes are shown for simplicity. A type of photodiode having different spectral characteristics by changing the thickness of the n-type diffusion layers 41, 42, and 43 is shown. A similar effect can be obtained by changing the thickness of the P ⁺ buried layers 44, 45, and 46. The point is that the position of the highest point of impurity concentration in the P ⁺ diffusion layers 44, 45, and 46 can be changed, and the effect is almost the same. As shown in the figure, the distances to this highest point are respectively x ₁ , x ₂ , and x ₃ .

The selection of x ₁ , x ₂ , x ₃ is arbitrary as long as they are different;
As long as it has three types of spectral characteristics including the three primary color components of red, green, and blue, the original color can always be reproduced by processing the output signal.

A typical example is x ₁ = 1 μm, x ₂ = 1.5 μm x ₃
= around 4 μm. In the n-type diffusion layer 41, electrons are mainly accumulated with blue light, in 42, the three primary color lights are almost evenly distributed (for brightness signals), and in 43, electrons are accumulated with red light contributing strongly. In fact, because the n-type diffusion layer 43 is thick, the effect that the holes of electron-hole pairs generated especially near the surface recombine within the n-type diffusion layer 43 and disappear together with the electrons becomes remarkable, and red light is emitted. This increases the proportion of electrons generated, resulting in a more preferable configuration. n
It goes without saying that if a blue filter can be placed on the type diffusion layer 41 and a red filter can be placed on the type diffusion layer 43, the spectral characteristics of each photodiode can be made more preferable. At this time, a perfect color filter may be placed on 42.

FIG. 7 shows an embodiment in which countermeasures against the blooming phenomenon are taken into consideration according to the sensitivities of the three origin components of red, green, and blue. As mentioned earlier, photodiodes using Si (including MOS diodes, of course) have high sensitivity to red light (long wavelength light) and low sensitivity to blue light (short wavelength light). Therefore, when each photodiode is set for red, green, and blue using a color filter, assuming that the junction capacitance of the photodiodes is the same, the red photodiode is most likely to cause blooming. is high, followed by the green photodiode, and the blue photodiode has almost no blooming phenomenon. on the other hand,
For example, if there is a P ⁺ buried layer 25 in Fig. 4 below the photodiode, the probability that electrons will disappear in that part is higher than in the case where the P ⁺ buried layer 25 does not exist. The amount of charge accumulated in the n-type diffusion layer across the junction is reduced. Therefore, for photodiodes with extremely low sensitivity such as blue photodiodes, it is better not to have a P ⁺ buried layer.

In FIG. 7, 50 is a P-type semiconductor substrate;
1, 52, 53 are n-type diffusion layers, 54, 55 are P ⁺
The buried layer or diffusion layer 56 is a SiO ₂ film. In this figure, only the photodiode is shown, and the gate electrode 23, the n-type diffusion layer (drain) 24,
A portion corresponding to the drain electrode (vertical output line) 27 is omitted. In this structure, the n-type diffusion layer 5
1, 52, and 53 constitute blue, green, and red photodiodes, respectively. Of course, the semiconductor substrate 5
Mosaic filters or stripe filters for blue, green, and red are formed on the SiO ₂ film 56 on the surface of 0, but are not shown. n
No P ⁺ buried layer is formed below the type diffusion layer 51. This is to keep the sensitivity of the blue photodiode as low as possible. Also, the P ⁺ buried layer 55 has a larger area in contact with the n-type diffusion layers 53 and 52 than the P ⁺ buried layer 54, but when the storage capacitance is the same, the red photodiode is larger than the green photodiode. Compared to a diode, there is a higher possibility that a blooming phenomenon will occur due to electrons overflowing from the n-type diffusion layer due to saturation of the storage capacitance. Therefore, the storage capacity of the red photodiode is increased. n-type region 57,
The power supply 58 is also one of the blooming prevention means. These two items do not need to be added. Between this n-type region 57 and p-type region 50
The method of applying a reverse bias to the pn junction to prevent blooming is detailed in Japanese Patent Application No. 1983-36163 previously filed by the present applicant. This blooming prevention means is also formed only on the red and green photodiodes, and is not formed below the blue photodiode. The reason for this is exactly the same as the reason mentioned above.

Note that when blue, green, and red photodiodes are formed on the same substrate, blooming prevention means are absolutely necessary for the red photodiode, not for the blue photodiode, and for the green photodiode. It is better to take preventive measures against this. Further, when the photodiodes are a cyan photodiode, a magenta photodiode, and a yellow photodiode, the cyan photodiode does not require blooming prevention means, but the other two do. Also, even if full-color photodiodes are present, anti-blooming means are necessary.

Although two blooming prevention means are described in the above embodiment, the invention is not limited to these in any way, and even when adding prevention means in the circuit structure, the blue or cyan photodiode may be used as a blooming prevention means. It is better not to install blooming prevention means, and it is absolutely necessary to install blooming prevention means for red, magenta, or yellow photodiodes.

Further, the manufacturing method of the structure shown in FIG. 7 will not be described in detail since it can be easily carried out by those skilled in the art, but is described in detail in, for example, Japanese Patent Application No. 144800/1983. Note that the n-type region 57 may be formed by selective ion implantation or diffusion from the back surface of the semiconductor substrate 50, or may be formed as a selectively buried layer in the semiconductor substrate. A modification of the embodiment of FIG. 7 is shown in FIG. In this embodiment, as in FIG. 7, the blue photodiode does not have a p ⁺ region as a blooming prevention means, and the green and red photodiodes do not have a p
A p ⁺ region 59 is formed. Of course, from the left, the photodiodes are for blue, green, and red.
The matters described in the previous embodiment also hold true in the same manner in this embodiment. In another embodiment, the P ⁺ diffusion layers 54 and 55 may be omitted in the embodiment shown in FIG.

It goes without saying that the present invention, which has been described in detail above, can be applied not only to MOS type solid-state imaging devices but also to CCD type and CID type.

The present invention, which has been described in detail above, can provide a highly sensitive color solid-state imaging device by adding no blooming prevention means to the green photodiode and adding blooming prevention means to the red photodiode.

[Brief explanation of the drawing]

Figure 1 is an explanatory diagram of a MOS solid-state imaging device, Figure 2
The figure shows the cross-sectional structure of one conventional picture element, Figure 3 is a diagram of the optical absorption characteristics of Si, Figures 4 to 6 are diagrams for explaining the present invention, and Figures 7 and 8. 1 is an embodiment diagram of the present invention.

Claims (1)

[Claims] 1 A semiconductor substrate of a first conductivity type, a plurality of semiconductor layers of a second conductivity type for photodiodes provided on this semiconductor substrate, a color filter provided corresponding to this semiconductor layer, and the semiconductor layer. a first conductive layer having a higher impurity concentration than the semiconductor substrate and covering a portion of each of a side surface and a bottom surface of the semiconductor layer from the side surface to the bottom surface; 2. A solid-state imaging device, wherein at least two of the semiconductor layers have a type impurity layer, and the contact area between the impurity layer and the semiconductor layer is different depending on the spectral characteristics of the color filter.