JPS6147031B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an image sensor having a photoelectric conversion function and a charge readout function.

2. Description of the Related Art Image sensors used in optical reading devices and the like are conventionally known to have a structure as shown in FIG. What is shown in the figure is a so-called CCD that distributes the accumulated charge in the light receiving section A to CCD (charge transfer device) shift register sections B ₁ and B ₂ on both sides and reads it out.
It is a line sensor. The light receiving section A and the shift register sections B ₁ and B ₂ each extend in a direction perpendicular to the drawing, and the light receiving section A forms a linear light receiving surface with an array of a plurality of MOS elements divided by meandering channel cuts. Furthermore, the shift register sections B ₁ and B ₂ are composed of a plurality of substrates arranged along a common channel and an array of MOS elements having an inverted buried layer.

The light receiving section A, which is the main body of the image sensor, is
For example, an N-type diffusion layer 2 is formed on a P-type silicon (Si) substrate 1, a PN junction J is formed between them, and a silicon oxide film (SiO ₂ ) is further formed above the PN junction of the substrate 1. A Hong gate electrode 4 made of polycrystalline silicon is provided through the electrode 3, and this electrode 4 is also covered with a silicon oxide film. Note that 5 ₁ , 5 ₂ in the oxide film 3
is a transfer gate electrode, 6 ₁ and 6 ₂ are transfer electrodes of the shift register section, and 7 is a light receiving section A.
This is a light shield electrode that covers the oxide film 3 except for only the upper part of the photogate electrode 4 in order to block incident light to other parts, and is made of, for example, an aluminum (Al) film.

The light receiving section A applies a voltage to the photogate electrode 4 to create a depletion layer in its lower semiconductor substrate, and stores charges generated by incident light (light image) from above the electrode 4 in the depletion layer. Photogate electrode 4
Since it is made of polycrystalline silicon, it has the disadvantage of poor sensitivity and wavelength characteristics. That is, this
A polycrystalline silicon film with a thickness of 2,000 to 3,000 Å is not completely transparent, and has the disadvantage that it is particularly difficult for light of short wavelengths to pass through.

In order to improve this point, an image sensor has been considered in which the photogate electrode 4 is removed from the structure shown in FIG. 1 and the oxide film 3 is formed above the PN junction J only. In this sensor, the transfer gates 5 ₁ and 5 ₂ are opened for charge accumulation, and the CCD
A voltage is applied to the electrodes 6 ₁ and 6 ₂ of the shift register section, and the N-type diffusion layer 2 of the light receiving section A is turned into a depletion layer by the voltage. That is, the electrodes 6 ₁ , 6 ₂ and 5 ₁ , 5 ₂
Furthermore, since this example is an N-channel device, when a positive voltage is applied, the N-type region 1a has a higher potential than the substrate 1. Moreover, since this region 1a communicates with the N-type diffusion layer 2, this layer 2 is made into a depletion layer, albeit incompletely, and its potential is increased. In this way, substantially the same effect as when the photogate electrode 4 is provided and a positive voltage is applied to it to create a charge storage state can be obtained. After that, electrode 6
Although the voltages ₁ , ₆₂ , ₅₁ , and ₅₂ are removed, the N-type diffusion layer 2 remains in the depleted state and high potential state described above. If a light image is projected in this state, charges corresponding to the light image will be generated and accumulated in the depletion layer and the PN junction J. If the light-receiving part is made up of only this oxide film 3, the wavelength characteristics will improve, but on the other hand, the depletion layer created by the CCD electrode is incomplete and the charge accumulation region is mainly the PN junction J, so the amount of charge accumulation is reduced and the light There is a drawback that the photoelectric conversion output relative to the input decreases. Furthermore, the silicon oxide film 3 is connected to the light shield electrode 7.
Since it is exposed through the opening of the silicon oxide film, it may cause instability of the characteristics of the silicon oxide film such as charge adhesion, and may even cause damage to the light receiving area A.
The problem is the deterioration of the characteristics.

The object of the present invention is to provide an image sensor that solves all of these problems at once, and includes a light receiving part that accumulates carriers generated by projecting light, and a light receiving part arranged along the light receiving part to accumulate carriers. In a CCD image sensor equipped with a charge transfer section, the light receiving section is connected to a semiconductor substrate having a PN junction, a silicon dioxide film formed on the substrate, a silicon dioxide film formed on the silicon dioxide film, and a conductivity higher than that of the silicon dioxide film. It is characterized by being made of a silicon nitride film that has good resistance and can be applied with a predetermined potential.

The silicon nitride film on the silicon dioxide film functions as a pseudo-photogate electrode by being maintained at an appropriate potential by the light shield electrode, etc., and forms a wide charge storage region in the semiconductor layer, reducing the amount of charge stored in the light receiving section. increase. Furthermore, when a silicon nitride film is used, the refractive index is approximately 2, whereas silicon dioxide is 1.5 and the surrounding air is 1. Therefore, light passes through each layer with a refractive index of 1, 2, and 1.5 through the silicon film, which has a refractive index of 4. It enters the substrate, passes through each layer with a refractive index of 1, 1.5, 4, and 1.5 in the case of Fig.
The attenuation (reflection) rate of incident light is lower than that of silicon substrates, and the sensitivity of the image sensor is improved over a wide range of wavelengths. Since the silicon dioxide film is covered with a silicon nitride film having some conductivity, there is no problem of charge accumulation and the surface of the oxide film is stabilized.

An embodiment of the present invention will be described below with reference to FIG. FIG. 2 is a sectional view showing an embodiment of the present invention applied to a CCD line sensor, and the same parts as in FIG. 1 are given the same reference numerals. In this embodiment, the light receiving part A' is the PN junction J of the semiconductor layer 1.
A silicon nitride film (Si ₃ N ₄ ) 8 with a thickness of about 1500 Å is formed above through a silicon oxide film (SiO) 3 with a thickness of about 1000 to 1500 Å, and the edge of the nitride film 8 is further covered with a conductive light shield. It is connected to the opening edge of an electrode (aluminum in this example) 7.

The refractive index of SiO ₂ with respect to air is about 1.5 as described above, whereas the refractive index of Si ₃ N ₄ is about 2, which is similar. Therefore, the nitride film 8 and the oxide film 3
The reflectivity of the incident light at the interface with the material is extremely small. Further, since the nitride film has high transparency like the oxide film, the attenuation of incident light is small, and the sensitivity, especially on the short wavelength side, is improved. Both the nitride film 8 and the oxide film 3 have remarkable insulating properties compared to ordinary conductive materials, but the nitride film 8 has a high ratio of silicon to nitrogen, specifically a mixture of silane and ammonia, during its vapor phase growth process. By changing the ratio, conductivity can be imparted and the conductivity can be adjusted within a certain range. Therefore,
The nitride film 8 can be held at zero potential by connecting it to the grounded light shield electrode 7, or at an appropriate positive or negative DC potential by separately providing an appropriate electrode. It functions in the same manner as the photogate electrode 4 in Figure 1, forming a wide accumulation region for charges generated by incident light in the semiconductor layer 1, and increasing the amount of charge accumulation in conjunction with the charge accumulation region of the PN junction J. can.

FIG. 3 is a plan view of a normal CCD line sensor, and the cross section taken along line x--x in FIG. 3 corresponds to FIGS. 1 and 2. The light-receiving section A' is a so-called photodiode array in which a plurality of light-receiving elements A'' (the above-mentioned diodes) separated by a meandering channel cut 9 are linearly arranged, and a photogate electrode 4 is provided on the light-receiving surface.
Alternatively, a nitride film 8 is provided, and the periphery is covered with a light shield electrode 7. The photoelectric conversion outputs (charges) of the light receiving sections A and A' are distributed and read out bit by bit as shown by white arrows A and B to the CCD shift register sections B ₁ and B ₂ on both sides. It is transferred in the direction of the white arrow C in the register sections B ₁ and B ₂ and then sent to the amplifier 1 via the output gates 10 ₁ and 10 ₂ .
1 ₁ , 11 ₂ , which is amplified and outputs 0 ₁ ,
Retrieved as 0 ₂ . Furthermore, the shift register section B ₁ ,
Prior to charge transfer by B ₂ , a voltage is applied to the transfer gate electrodes 5 ₁ and 5 ₂ to open the gates and transfer charges all at once from the light receiving section A' to the CCD register section. Since the serial charge transfer by the CCD electrodes 6 ₁ and 6 ₂ is similar to a normal CCD driving method, the details thereof will be omitted.

As is clear from the above, the image sensor of the present invention has improved wavelength characteristics and a large amount of accumulated charge, so it is possible to solve the contradictory problems of conventional image sensors all at once. This has the advantage that passivation can be performed at the same time.

[Brief explanation of the drawing]

FIG. 1 is a sectional view showing an example of a conventional CCD image sensor, FIG. 2 is a sectional view showing one embodiment of the present invention, and FIG. 3 is a plan view of a CCD line sensor. 1... P-type semiconductor layer, 2... N-type diffusion layer, 3...
... silicon oxide film, 5 ₁ , 5 ₂ ... transfer gate electrode, 6 ₁ , 6 ₂ ... CCD electrode, 7 ... light shield electrode, 8 ... silicon nitride film (insulating film), 9 ... channel cut, 10 ₁ , 10 ₂ ...
...output gate, 11 ₁ , 11 ₂ ... amplifier,
A′... Light receiving section, B ₁ , B ₂ ... CCD shift register section.

Claims (1)

[Claims] 1. A CCD image sensor comprising a light receiving section that accumulates carriers generated by projecting light and a charge transfer section that is disposed along the light receiving section and transfers the accumulated carriers. A semiconductor substrate having a PN junction, a silicon dioxide film formed on the substrate, and a silicon nitride film formed on the silicon dioxide film, which has better conductivity than the silicon dioxide film and can be applied with a predetermined potential. Featured image sensor. 2. The image sensor according to claim 1, wherein the silicon nitride film is connected to a light shield electrode that covers a surface other than a light receiving part such as a charge transfer part.