JPH054877B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to a solid-state imaging device that can perform color imaging with a single-chip configuration without using a color separation filter (hereinafter abbreviated as a color filter). Up until now, color solid-state imaging devices have been developed with a focus on single-chip configurations, which are most advantageous for miniaturization, and methods have been adopted in which color filters are arranged in a mosaic pattern to match the pixels. ing.
However, with this method, the following problems remain that are difficult to solve in terms of reliability and performance. (1) The color filters currently in use are organic filters composed of gelatin films, so they are heat resistant and
Due to insufficient light resistance, it does not have the same long lifespan as semiconductors. (2) Since the manufacturing precision of color filters is much lower than that of semiconductors, it becomes extremely difficult to manufacture color filters as solid-state image sensors become smaller and pixel dimensions become smaller due to an increase in the number of pixels (higher resolution). (3) Because the three primary colors (or multiple colors including complementary colors) necessary for color reproduction are arranged in a mosaic pattern, one color appears in every two or three pixels, resulting in extremely poor color resolution. Enough is enough. The resolution of problem (3) above is the most important performance of a camera, and is an element that greatly influences image quality. Generally, the resolution of a solid-state image sensor is determined by the number of photodiodes (photoelectric conversion units such as photodiodes) that serve as pixels, but currently the resolution is approximately 500 in the vertical direction and 400 in the horizontal direction in a 2/3-inch light receiving area. An individual is a typical example. However, in order to achieve resolution comparable to color negative film with this solid-state image sensor, as long as conventional mosaic color filters are used, the number of pixels must be at least 2,200 vertically and 2,200 horizontally, or 2/3 inch. The light-receiving area requires 5 million pixels. (Actually, the r, m, s granularity of the film must be taken into account, and the value will be 20 to 30 times more.) When the number of pixels is as above, a cell containing one pixel The size is approximately 3 μm in the vertical direction and approximately 4 μm in the horizontal direction (in the case of a 2/3-inch light receiving area), so the manufacturing precision of the color filter requires the precision of a VLSI, which is extremely difficult. At this time, if all the color signals necessary for color reproduction can be obtained from one pixel, in order to achieve a resolution comparable to that of color negative film, approximately 1,300 signals in the vertical direction and approximately 1,300 signals in the horizontal direction with a 2/3-inch light-receiving area can be obtained. 1,300 pixels, or 1.7 million pixels, would be sufficient. In this case, the cell size including one pixel is approximately 5.1 μm in the vertical direction and approximately 6.8 μm in the horizontal direction.
It can be seen that the manufacturing dimensions of the solid-state image sensing device are also much easier. However, although solid-state imaging devices that read out multiple color signals simultaneously from one pixel have been proposed so far (for example, see Japanese Patent Application Laid-Open No. 118932/1983),
It is far from a practical level. This is because it is extremely difficult to simultaneously realize means for obtaining a plurality of color signals and means for reading them as two-dimensional information. An object of the present invention is to realize a new solid-state imaging device that is equipped with means for obtaining a plurality of color signals necessary for color reproduction from a photoelectric conversion section of one pixel, and means for simultaneously reading out each color signal as two-dimensional information. It is in. In the configuration of the present invention, in order to obtain a plurality of color signals,
The photoelectric conversion section is formed in layers in the thickness direction perpendicular to the substrate, and multiple resistive gate electrodes (hereinafter abbreviated as RG electrodes) are embedded in order to simultaneously read out color signals from each layer.
It consists of a signal readout section mainly composed of. First, let's consider the photoelectric conversion section. Generally p
-Spectral sensitivity characteristic R(λ) of n-junction photodiode
R(λ) = e・η(λ)/(hc/λ) ・U(λ _G −λ) …(1) However, Eg: band gap of the intrinsic semiconductor λ _G = hc/Eg: cutoff for the band gap Eg Wavelength η (λ): Effective quantum efficiency U (λ _G - λ) = 1 (λ≦λ _G ) = 0 (λ > λ _G ) In addition, the quantum efficiency η (λ) is the absorption coefficient α (λ),
Using the diffusion depth xj, it can be approximately expressed by the following equation. η(λ)1−exp{−α(λ)·xj} (2) From equations (1) and (2), the following two methods can be considered to determine the wavelength selectivity of the photoelectric conversion section. () Utilizes the spectral sensitivity characteristic determined by the diffusion depth xj of the pn junction. () Utilizes the spectral sensitivity characteristics of a heterojunction composed of semiconductors with different Eg. Before considering the influence of the spectral sensitivity characteristics of the diffusion depth xj shown in (), let us consider Si (silicon) as a typical example.
The spectral sensitivity characteristics of the spectral sensitivity are as shown in Fig. 1. In contrast to the ideal spectral sensitivity characteristic represented by the broken line, in reality, the short wavelength side (λ<400 nm) is attenuated due to surface traps, and the long wavelength side (λ>800 nm) is attenuated due to bulk traps. FIG. 2 shows how such spectral sensitivity characteristics change with the diffusion depth xj. From FIG. 2, it can be seen that as xj becomes smaller, there is an improvement on the short wavelength side. Therefore, the diffusion depth of the photoelectric conversion section formed in layers
By utilizing the difference in xj, a plurality of different color signals can be obtained. Next, in order to consider the influence of Eg on the heterojunction shown in (), when formula (1) is applied to the case of a semiconductor heterojunction with Eg ₁ Eg ₂ , the following formula is obtained. R (λ) = e・η (λ) / (hc / λ) ・[U (λ _G1 - λ) + U (λ _G2 - λ) ... (3) However, λ _G1 = hc / Eg ₁ < λ _G2 = hc/Eg ₂ U(λ _G1 −λ)=1(λ≦λ _G1 ) =0(λ>λ _G1 ) U(λ _G2 −λ)=1(λ≦λ _G2 )=0(λ>λ _G2 ) The bandgap of the semiconductor on the surface where light enters is
Eg ₁ , when the wavelength λ of light is λ>λ _G1 , U(λ _G1 - λ)=0, which means that λ>
λ This indicates that the displayed semiconductor becomes transparent with respect to the wavelength λ of light of _G1 . (This is generally called the window effect)
This means that when photoelectric conversion is performed using a heterojunction as shown in FIG. 3a, spectral sensitivity characteristics as shown in FIG. 3b can be obtained. Therefore, by using a combination of dongap Eg of photoelectric conversion sections formed in layers, a plurality of different color signals can be obtained. Next, we will consider the resistive gate (RG) electrode used in the signal readout section. The basic structure of the resistive gate electrode is shown in FIG. 4a. n ^- region 402 formed on p substrate 401
A resistive gate electrode 4 is placed on the resistive gate electrode 4 with an insulating film 403 thereon.
04 is created and a voltage E is applied across it, a potential distribution of the n ^- region 402 as shown in FIG. 4b is obtained. At this time, n ^- area 40
Assuming that the depletion potential of 2 is ψ=ψ ₀ , the range indicated by A corresponds to the region to be depleted. To realize the scanning function using such resistive gate electrodes, at least two resistive gate electrodes are required as shown in FIG. 5a. FIGS. 5a and 5b show a top view and a sectional view, respectively, and FIG. 5c shows the potential distribution. Resistive gate electrodes 504 and 505 are formed on an n ^- region 502 formed on a p-substrate 501 with an insulating film 503 interposed therebetween. When a voltage E ₁ is applied between terminals B and C at both ends of the resistive gate electrode 504, the potential distribution becomes ψ ₁ , and when a voltage E ₂ is applied between terminals D and E at both ends of the resistive gate electrode 505. Then, the potential distribution becomes ψ ₂ . At this time, the depletion potential of the n ^-region 502 is ψ=ψ ₀
Then, for the resistive gate electrode 504, the range indicated by F becomes the depletion region, and for the resistive gate electrode 505, the range indicated by G becomes the depletion region. Therefore, when looking at the n ^- region 502 in a direction perpendicular to the resistive gate electrodes 504 and 505, a channel (the hatched portion of the n ^- region 502 in FIG. 5b) is formed in the range indicated by H with the width. That will happen. To scan such a channel, for example, to the right, the sawtooth waves R ₁ and R ₂ are applied between the B and C terminals, respectively.
It may be supplied between the D and E terminals via a capacitor. The details of the present invention, which is composed of a photoelectric conversion section for obtaining a plurality of color signals and a signal readout section using a resistive gate electrode for reading out a plurality of color signals, as described above will be described in detail in an embodiment. Explain using. FIG. 6 shows one embodiment of the photoelectric conversion section of the present invention.
a top view, b X-X′ cross-sectional view regarding each pixel,
c shows a Y-Y' cross-sectional view. Since this pixel has continuous parts in the Y-Y' direction (for example, n ⁺ areas 615, 609, 603, etc.), the solid-state image sensor can It is constructed by arranging them in the −X′ direction. FIG. 6 shows an n ⁺ region 603 separated by an insulator 601 on a p-substrate 602, and n ⁺ type resistive gate electrodes 604 and 60 buried on the n ⁺ region 603.
5, an n ⁺ region 607 separated by an insulator 601 on the p region 606, and an n ⁺ region 6
07, an n ⁺ region 609 separated by an insulator 601 above the p ⁺ region 608, and an ^{n+ type} resistive gate electrode 610, 611 buried above the n ⁺ region 609. A p region 612, an n+ region 613 separated by an insulator 601 on the p region 612, a p+ region 614 on the n ⁺ region 613, a p ⁺ region 614 on the n ⁺ region 613 ^,
n ⁺ separated by an insulator 601 on top of the region 614
region 615, embedded above n ⁺ region 615
p including n ⁺ type resistive gate electrodes 616, 617
A pixel structure is shown in which a region 618, an n ⁺ region 619 separated by an insulator 601 on the p region 618, and a protective film 620 on the n ⁺ region 619 are formed in this order. The unit pixels shown in Fig. 6 are B, G,
It is divided into three areas of R, and three types of color signals can be read out. In region B, the n ⁺ region 619 is the main photoelectric conversion section, and performs photoelectric conversion in response to light with wavelengths including 400 nm to 550 nm (equivalent to blue light), and as a result,
The generated signal charge is transferred to the resistive gate electrodes 616, 61.
n ⁺ area 619 to n ⁺ area 6 when scanned with 7
n ⁺ corresponding to the signal line SL _B through the channel (indicated by a dotted line) in the thickness direction formed toward 15
The signal is read out to area 615 as a first color signal. In the G region, the n ⁺ region 613 is the main photoelectric conversion section, and performs photoelectric conversion in response to light with wavelengths including 500 nm to 600 nm (equivalent to green light), and the resulting signal charge is resistive. gate electrode 610,
A second color signal is transmitted to the n ⁾ area 609 corresponding to the signal line SL _G through a channel (indicated by a dotted line) in the thickness direction formed from the n+ area 613 to the n ⁺ area 609 when the image is scanned by the n+ area 611. It is read as . In the R region, the n ⁺ region 607 is the main photoelectric conversion section, which performs photoelectric conversion in response to light with wavelengths ranging from 550 nm to 700 nm (equivalent to red light), and the resulting signal charge is transferred to the resistive gate electrode. 604,605
n ⁺ area 607 to n ⁺ area 60 when scanned with
The third color signal is read out to the n ⁺ region 603 corresponding to the signal line _SLR through a channel (indicated by a dotted line) in the thickness direction formed toward the third color signal. In addition,
P regions 602, 606, 608, 61 at ground potential
A positive voltage is applied to the n ⁺ regions 603, 609, and 615, which serve as signal lines for the signals 2, 614, and 618. As a specific method for obtaining the above-mentioned first to third color signals, () Utilize the difference in spectral characteristics depending on the diffusion depth xj. () Utilizes the difference in spectral characteristics of heterojunctions with different band gap energies Eg. I have already mentioned that two of these are possible. For the unit pixel in Figure 6, the diffusion depth of ()
A specific example of applying the effect of xj will be described below. If the unit pixel in Fig. 6 is composed of only Si (silicon), the area B should be 1 μm with reference to Fig. 2.
If the G region is formed in the depth direction within 2 μm, and the R region is formed in the depth direction within 4 μm, the first color signal C ₁ obtained in the B region is obtained.
corresponds to all of blue light b, green light g, and red light r, so C ₁ = r + g + b ... (4) Similarly, the second color signal C ₂ obtained in the G region is C ₂ = r + g ... ( 5) Similarly, the third color signal C ₃ obtained in the R region is also calculated as follows: C ₃ = r...(6) From equations (4), (5), and (6), g = C ₂ - _{C 3} , b= _C1 − _C2 ,
Since r=C ₃ and color signals corresponding to the three primary colors are obtained, the first to third color signals obtained by the unit pixel shown in FIG.
It can be seen that color reproduction is possible by using the color signal. As described above, it has become clear that a plurality of color signals necessary for color reproduction can be simultaneously obtained from one pixel by utilizing changes in spectral characteristics due to differences in diffusion depth xj. Since the pixel portion in this case can be formed using only the same semiconductor as the substrate, it is easy to manufacture using a process based on ordinary epitaxial crystal growth technology, and has the characteristics that it is easy to obtain a clean interface. However, theoretically, equation (2), qualitatively,
As can be seen from equations (4) and (5), it is difficult to obtain a plurality of color signals necessary for color reproduction in a narrow wavelength range, resulting in the problem that signal processing is required. This problem can be solved by utilizing the change in spectral characteristics obtained by heterojunctions with different band gap energies Eg (), and a specific example will be described below. First, the basic principle will be explained using FIG.
Figure 3a shows a heterojunction photodiode with a p ⁺ nn ⁺ structure, with a p ⁺ region 301 serving as the top contact.
The n ⁺ substrate 303 is formed of a - group binary compound InP, and the n region 302, which becomes a drift region (depletion region), is a - group ternary compound.
Made of GaAsSb. If Eg is _the bandgap energy (hereinafter abbreviated as Eg) of InP, the cutoff wavelength λ _G1 is λ _G1 = hc/Eg ₁ .
Further, if Eg of GaAsSb is Eg ₂ , its cutoff wavelength λ _G2 is given by λ _G2 =hc/Eg ₂ . At this time, since Eg ₁ >Eg ₂ , the wavelength λ of the light corresponding to the color signal becomes λ _G1 <λ<λ _G2 as shown in FIG. 3b. Further, it is also possible to use a - group quaternary compound (for example, InGgAsP) for the n region 302. In addition, a method called compositional grading is also used to move away from the limitations of a fixed molar ratio of multiple components and a corresponding fixed cutoff wavelength and to increase the degree of freedom in setting. (This also helps to get a clean interface)
This is similar to the top contact p ⁺ region 702, as in the heterojunction photodiode shown in FIG.
The n ⁺ substrate 701 is InP, and the drift region n
When the mold region 703 is In _x Ga _1-x As _Y P _1-Y , the mixing gradient is In _x Ga _1-x as in the regions 704 and 705.
As _y P _1-y, where it touches InP, x=
1, y=0, and the molar ratio is made with a gradient so that _X = _x , y= _Y where it contacts _In How to create a gradient is the parameters x, y
Generally, it changes continuously or in steps. Of course, regardless of the mixing gradient, In _x Ga _1-x AS _Y P _1-Y
In some cases, it is sufficient to use just one, and you can use it depending on the purpose. As a specific example of applying the above method to the unit pixel in FIG. 6, blue light is applied to the B area, green light is applied to the G area, and R
A case in which red light is assigned to an area will be explained. For simplicity, λ _G1 400nm (corresponds to E _G1 3eV),
λ _G2 500nm (compatible with E _G2 2.4eV), λ _G3 600nm
(corresponding to E _G3 2eV), λ _G4 700nm (corresponding to E _G4 1.7eV), the first color signal C _B in the B region, the second color signal HC _G in the G region, and the third color signal in the R region Consider the case where the wavelengths to which _CR responds are C _B ; λ _G1 to λ _G2 , C _G ; λ _G2 to λ _G3 , and _CR ; λ _G3 to _{λ G4} . In order to allocate the cutoff wavelengths λ _G1 to λ _G4 in FIG. 6 to satisfy the above conditions, for example, the n ⁺ region 6
19, λ _G2 in the P region 618, λ _G3 in the n ⁺ region 613, and λ _G4 in the n ⁺ _region 607. Even without λ _G1 in the B region, the short wavelength response should be attenuated by the surface states, as shown in Figure 1.
It is also possible to allocate λ _G2 to the n ⁺ region 619, λ _G3 to the n ⁺ region 613, and λ _G4 to the n ⁺ region 607. Note that the allocation of λ _G2 to λ _G4 may be formed using a mixed gradient. For parts other than those specified, λ _G1 or λ _G2 may be assigned. In particular, the n ⁺ regions 615, 609, 603 for signal transmission,
If it is formed of a compound semiconductor that exhibits a window effect with respect to the wavelength to be photoelectrically converted, there is no need to worry about generation of false signals. Next, the specific materials to be used will be explained. FIG. 8 shows the relationship between Eg and lattice constant of - group multicomponents. From FIG. 8, we cannot find a compound semiconductor with uniform lattice constants corresponding to λ _G1 to λ _G4 above, but
In compound semiconductors corresponding to λ _G2 to λ _G4 , it is possible to make the lattice constants uniform. The more uniform the lattice constants are, the better the interface properties will be, and the more ideal the device properties will be. For example, AlP _x Sb _1-x for λ _G2 and Ga _y for λ _G3
Ga _y ′Al _{1- y ′As may be used for Al 1-y} As and λ _G4 .
At this time, each part in FIG. 6 includes a p substrate 602, p regions 606, 608, 612, 614, 618, and
n ⁺ type resistive gate electrodes 604, 605, 610,
611, 616, 617, further n ⁺ area 603,
609,615,619 becomes AlP _x Sb _1-x , n ⁺
The region 613 becomes Ga _y Al _1-y As, and the n ⁺ region 607
becomes Ga _y ′Al _1-y ′As. For actual manufacturing, MBE
Epitaxial crystal growth technology using the (Molecular Beam Epitaxy) method may be used. Note that the band gap energy Eg can be set using the above-mentioned compound semiconductor, but it can be slightly corrected by doping with impurities.
The impurities used in _the − group compound ^are as shown _in the table below.
Li, Se, Te, etc. are used, and Li, Sn, Pb, O, S, Se, Te, etc. can be used to make Ga _y Al _1-y As n ⁺ type. At this time, when the impurity doping becomes high, a state called degenerate doping a or impurity tail b can be realized as shown in FIG.

【table】

[Table] From FIG. 9, it can be seen that to increase Eg, it is sufficient to achieve a degenerate doping state, and to decrease Eg, it is sufficient to achieve an impurity tail state. Up to now, it has been described that the unit pixel in FIG. 6 is constructed from a - group multi-element compound, but it is also possible to use a - group multi-element compound. FIG. 10 shows the relationship between Eg and lattice constant of - group multicomponent compounds. As an example, in the n ⁺ region 619 of _FIG.
500nm (corresponding to E _G2 2.4eV), λ _G3 in n ⁺ region 613
600nm (corresponding to E _G3 2eV), λ _G4 in n ⁺ region 607
Considering the case of allocating 700nm (corresponding to E _G4 1.7eV), from Fig. 10, λ _G2 has Zn _x Cd _1-x S _y
Te _1-y , λ _G3 has Zn _x ′Cd _1-x ′S _y Te _1-y , λ _G4 has
Zn _x ″Cd _1-x ″Te may be used. However, since the crystal structure changes from zincblende to wurtzite as the molar ratio changes, it is preferable to use a mixing gradient. As mentioned above, it has become clear that multiple color signals necessary for color reproduction can be simultaneously read out from one pixel even by utilizing changes in spectral characteristics determined by heterojunctions with different Eg. Next, a solid-state image sensing device constructed using a photoelectric conversion section as shown in FIG. 6 will be explained using a specific example. FIG. 11 shows the cross-sectional structure of a unit pixel and its equivalent circuit. FIG. 11a is the same as FIG. 6b, and shows the state of application of the bias voltage. FIG. 11b is an equivalent circuit of a unit pixel, where photodiode DB ₁ corresponds to a p-n ⁺ junction photodiode of n ⁺ region 619 and p region 618, and photodiode DB ₂ corresponds to n ⁺ region 613 and p region. 612, and photodiode DB ₃ corresponds to the pn ⁺ junction photodiodes of n ⁺ region 607 and p region ⁶⁰⁶ . Furthermore, the resistive gate electrodes GB ₁ , GB′ ₁ correspond to the n ⁺ gate regions 617, 616, the resistive gate electrodes GG ₁ , GG′ ₁ correspond to the n ⁺ gate regions 611, 610, and the resistive gate electrodes GR ₁ and GR' ₁ correspond to n ⁺ gate regions 605 and 604. Furthermore, the signal line LB corresponds to the n ⁺ region 615 to which the voltage E _B was applied, and the signal line LG corresponds to the n ⁺ region 615 to which the voltage E _G was applied.
9, the signal line LR is n ⁺ to which the voltage E _R is applied.
Corresponds to area 603. The signal from the photodiode DB ₁ is read out to the signal line LB through a channel formed where the hatching portions of the two resistive gate electrodes GB ₁ and GB' ₁ overlap, and is detected by the load resistor R _LB. Ru. Similarly, the signals of photodiodes DG ₁ and DR ₁ are connected to load resistors R _LG ,
Detected by R _LR . FIG. 12 shows an embodiment of a solid-state imaging device constructed using the equivalent circuit shown in FIG. 11b. In FIG. 12, the unit pixels of FIG. 11 are arranged in a matrix of m columns in the horizontal direction and n rows in the vertical direction. The resistive gates GB ₁ and GB' ₁ are used to scan the signals of m photodiodes DB ₁ l to DB ₁ m arranged in the horizontal direction. Similarly, the resistive gate electrodes GG ₁ , GG' ₁ are composed of m photodiodes.
It is used to scan the signals from DG ₁ l to DG ₁ m, and the resistive gate electrodes GR ₁ and GR' ₁ are used to scan the signals from m photodiodes DR ₁ l to DR ₁ m. Note that terminals A 1 and A 1 ' of resistive gate electrodes GB ₁ , GG ₁ , and GR ₁ are commonly connected, and resistive gate electrodes GB ₁ , GG ₁ , and GR 1 are
GB′ ₁ , GG′ ₁ , and GR′ ₁ are commonly connected to terminals B ₁ and B′ ₁ . Furthermore, a voltage E ₁ is applied between the terminals A ₁ and A′ ₁ in order to read out the signal of the photodiode, a sawtooth wave RO ₁ is applied via the capacitance, and the terminals B ₁ and
A voltage E ₂ is applied across B′ ₁ and a sawtooth wave RO ₂ is applied through the capacitance to perform horizontal scanning, and the signals from the photodiodes DB ₁ l to DB ₁ m are transferred to the signal line Read out to LB ₁ , photodiode
The signals from DG ₁ l to DG ₁ m are read out to signal line LG ₁ ,
The signal from photodiode DR ₁ l to DR ₁ m is the signal line
Read to LR ₁ . This is connected to the terminals (A ₁ , A' ₁ ) ~ (An, A'n) and (B ₁ , B' ₁ ) ~ of the resistive gate electrodes arranged in n rows.
(Bn to B'n) simultaneously, signals are simultaneously read out from all signal lines (LB ₁ to LBn), (LG ₁ to LGn), and (LR ₁ to LRn) in parallel. If scanning is also required in the vertical direction, a vertical scanning circuit may be introduced. Specifically, as shown in FIG. 13, pulse transmission lines X ₁ , X ₂ , . . . are sequentially connected to the vertical scanning circuit 1401.
_{By applying a pulse to _ _ _}
，
Vertical scanning is performed by sequentially turning on (Qn, Q′n, Pn, Pn′). As described above, according to the present invention, a novel method is provided which includes means for obtaining a plurality of color signals necessary for color reproduction from a photoelectric conversion section of one pixel, and means for simultaneously reading out the signals for each color as two-dimensional information. A solid-state imaging device can be realized. Furthermore, although the present invention has been described with reference to an embodiment in which three photoelectric conversion sections are provided in the thickness direction, the number of photoelectric conversion sections may be increased depending on the application, and the spectral characteristics may be divided into fine sections and assigned. Furthermore, it goes without saying that it can be used purely as an image sensor for detecting three-dimensional information without changing the spectral characteristics of the photoelectric conversion parts arranged in the thickness direction; in this case, there is no need to consider the spectral characteristics, so , a solid-state image sensor can be realized. The present invention described above is capable of obtaining multiple color signals necessary for color reproduction from one pixel.
It has the following effects as a color solid-state image sensor. (1) Eliminates the need for thermally unstable organic (gelatin) filters for color separation. (2) Color signals can be read out independently, increasing the degree of freedom in colorization methods. (3) High reliability can be achieved because the color image sensor can be constructed using only semiconductor materials. (4) Since the resolution in color is exactly the same as in black and white, a significantly higher resolution can be achieved with the same number of pixels as a conventional solid-state image sensor. Specifically, when colorizing using a conventional color filter with 1300 pixels horizontally x 1300 pixels vertically, 2200 pixels horizontally x 1300 pixels vertically.
Equivalent to 2200 vertical pixels. Furthermore, since signal reading is performed using a combination of resistive gate electrodes, the following effects are achieved. (5) High-frequency clock pulses for horizontal scanning are not required.
Horizontal scanning is possible with a bias power supply and a low frequency sawtooth wave. (6) It is possible to read out signals in parallel from all rows without using a vertical scanning circuit, and it can be used as a high-speed readout image sensor for information processing. This can exhibit sufficient performance in fields such as electronic still cameras, color solid-state cameras for broadcasting, and holographic information detection cameras for real-time processing, and can realize solid-state imaging devices that surpass image pickup tubes.

[Brief explanation of the drawing]

Figure 1 is a characteristic diagram showing the theoretical and experimental curves of the spectral characteristics of a Si p-n junction photodiode.
The figure shows the relationship between the spectral characteristics and the diffusion length xj of a Si p-n junction photodiode.
Figure 4 is a diagram showing the structure a and spectral characteristics b of a heterojunction photodiode made of elemental and ternary - group compound semiconductors. Figure 4 is a cross-sectional view a and potential distribution b of a basic resistive gate electrode. The figure shows a top view a, a cross-sectional view b, and a diagram showing the potential distribution c of a scanning circuit using two resistive gate electrodes.
6a, b, and c are top views and sectional views taken along line X-X' of a unit pixel of a solid-state imaging device according to an embodiment of the present invention;
Y-Y' cross-sectional view, Figure 7 is a diagram showing the cross-sectional structure of a heterojunction photodiode with a mixed gradient, and Figure 8 is a characteristic diagram showing the relationship between bandgap energy Eg and lattice constant of - group compound semiconductor. ,
Figure 9 shows how Eg changes in a highly impurity state, and shows a state called degenerate doping a and impurity tail b. Figure 10 shows the relationship between band gap energy Eg and lattice constant of a compound semiconductor. FIG. 11 is a diagram showing an equivalent circuit of a unit pixel of a solid-state imaging device according to an embodiment of the present invention, and FIG. 12 is a diagram showing a solid-state imaging device configured using a unit pixel according to an embodiment of the present invention. FIG. 13 is a circuit diagram showing an embodiment of the solid-state imaging device of the present invention having a vertical scanning function. 619,613,607...n ⁺ area (photoelectric conversion section), 615,609,603...n ⁺ area (signal transmission means), 617,616,611,610,
605, 604...n ⁺ area (scanning area).

Claims (1)

[Scope of Claims] 1. A photoelectric conversion section of one conductivity type formed corresponding to each pixel region allocated in a matrix on a semiconductor substrate of one conductivity type, and a photoelectric conversion section of the other conductivity type formed corresponding to each pixel region allocated in a matrix on a semiconductor substrate, An image pickup device consisting of a pair of upper and lower resistive gate portions of the other conductivity type formed separately below the photoelectric conversion and a signal line portion of the other conductivity type formed separately below the resistive gate is mounted on the semiconductor substrate. A solid-state imaging device characterized in that three pieces are stacked one on top of the other in the depth direction. 2 Each pixel region allocated in a matrix on a semiconductor substrate of one conductivity type is arranged on the first region except for a first region of the other conductivity type which becomes a signal line and a central part of the pixel region. A second region of one conductivity type serving as a readout region in which a pair of upper and lower resistive gates of the other conductivity type is embedded, and a third region of the other conductivity type forming a photoelectric conversion section formed thereon. A solid-state imaging device in which a plurality of pixel portions are formed so as to be stacked in the depth direction of a semiconductor substrate, with a fourth region of one conductivity type sandwiched between a set of pixel portions.