JPH03253082A - Photoelectric conversion device - Google Patents

Abstract

PURPOSE:To easily obtain a large area of low noise and high sensitivity, by clamping a light absorbing layer having forbidden bandwidth of Eg1 and a step-back structure between electron injection blocking layers, in which structure a forbidden bandwidth continuously changes between the minimum forbidden bandwidth Eg2 and the maximum forbidden bandwidth Eg3. CONSTITUTION:A light absorbing layer 310 which has a forbidden bandwidth Eg1 and absorbs light and a plurality of step-back structure layers 301, 303, 305, 307, 309 are sandwiched between a P-type semiconductor layer 311 and an N-type semiconductor layer 315 which turn to electron injection blocking layers. In the step-back structure layer, a forbidden bandwidth which multiplies carriers generated by absorbing light continuously changes between the minimum forbidden bandwidth Eg2 and the maximum forbidden bandwidth Eg3. The P-type semiconductor layer 311 and an electrode 313, and the N-type semiconductor layer 315 and an electrode 314 are electrically connected and formed on a glass substrate 316. Thereby a large area which is excellent in high speed response and has low noise properties and high sensitivity for visible light can be easily obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a photoelectric conversion device, and particularly to a photoelectric conversion device that utilizes the avalanche effect.

INDUSTRIAL APPLICATION This invention is suitably used for the photometry sensor of a camera, the image sensor for image reading devices, such as a facsimile machine and a copying machine, or the light receiving sensor of optical communication devices, etc.

[Prior art] Video information systems that use light as a medium for information signals, optical communications,
In other industrial and consumer fields, semiconductor light-receiving elements that convert optical signals into electrical signals are one of the most important and fundamental components, and many of them have already been put into practical use. In general, a semiconductor photodetector is required to have a high signal-to-noise ratio for its photoelectric conversion characteristics.

Among these, an avalanche photodiode (hereinafter referred to as APD) that utilizes the avalanche effect has a high gain and a fast response speed, and is therefore a promising candidate for a semiconductor light-receiving device that satisfies these requirements.

Many APDs are currently in practical use, especially as semiconductor light-receiving elements in optical communication systems, using compound semiconductors such as InGaAs, and have further improved the basic performance of the elements, such as low noise, high-speed response, and high gain. Development is progressing, and application to other fields, such as visible light receiving elements, is also desired.

FIG. 42 is a longitudinal sectional view showing the structure of a conventional APD for optical communication.

In the figure, 101 is an n'' type InP layer, 102 is an n
103 is an n-type InP layer, 104 is a p-type InGaAs layer, and 103 is an n-type InP layer.
It is a + type InP layer. Here, the n-type InGaAs layer 10
2. The n-type InP layer 103 and the p-type InP layer 104 are formed in a mesa shape. A p-electrode 106 is formed on the upper surface of the p-type 1nP layer 104, leaving a window 105, and an n-type InP layer 104 is formed.
An n electrode 107 is formed on the back surface of the nP layer 101. 10
8 is a passivation film. Here, the p electrode 106 and the n electrode 107 are biased in opposite directions, and the window 105
When light is irradiated from the substrate, the light is absorbed by the n-type InGaAs layer 102 (which becomes a light absorption layer), and light-to-electricity conversion is performed.

That is, electron-hole pairs formed in the n-type InGaAs layer 102 travel toward the n-electrode 107 and the p-electrode 106, respectively. Since n-type InP #103 (which becomes the multiplication layer) has a strong electric field, a sagging phenomenon occurs in which many electron-hole pairs are formed during the hole traveling process, and multiple electron-hole pairs are formed for one photon. A multiplication effect occurs in which electron-hole pairs are formed.

As a result, even weak incident light can be detected. However, in the conventional structure, the practical multiplication is as small as about 2, and due to fluctuations inherent in the multiplication process, excessive multiplication noise occurs and the signal-to-noise ratio decreases.

In view of these two drawbacks, as an APD for optical communication, for example, F. Capasso et al.
Publications and IEEE Electron Device
Letters 3rd EDL edition (71-7 of 19821)
On page 3, the authors propose a low-noise APD that can be used in optical communication systems that are made using compound semiconductors mainly belonging to the ■-v group using molecular beam epitaxy (MBE).

The element has a composition ratio of its constituent materials (for example, nl −
If a compound semiconductor belonging to group V is used as its constituent material, the semiconductor layer has a band gap that is continuously changed from a narrow side to a wide side by changing the composition ratio of a group (■) semiconductor and a group (■) semiconductor. It features a multilayer heterojunction structure that promotes ionization by stacking a large number of layers and utilizing the step-like transition region (hereinafter abbreviated as step-back structure) of the energy band formed at that time. The schematic structure of the proposed device will be explained using FIGS. 43(a) to 43(c).

FIG. 43(a) is a longitudinal cross-sectional view of this device, in which a step-back structure layer 201 consisting of five layers serving as a multiplication layer,
203, 205, 207 and 209 are light absorption layers n
The electrode 213 is in ohmic contact with the n-type semiconductor layer 211, and the electrode 214 is in ohmic contact with the n-type semiconductor layer 215.

FIG. 43(b) is a structural diagram of the energy band of the bandgap gradient layer in the non-bias state of this device, and three bandgap gradient layers are shown. Each layer has a composition that changes the bandgap linearly from a narrow bandgap Eg2 to a wide bandgap Eg3.

The step-back magnitudes of the conduction band and valence band are indicated by ΔEc and ΔEv, respectively. As will be explained later, △
Ec is set larger than ΔEv.

FIG. 43(C) is a structural diagram of an energy band when a reverse bias voltage is applied to this element. Note that the reverse bias voltage does not need to be a strong electric field compared to the APD shown in FIG. 42 described above.

Here, when light enters from the p-type semiconductor layer 211, the light absorbed by the p-type semiconductor layer and each step-back structure layer is
Photoelectric conversion is performed in the same manner as in the APD described above, and the formed electron-hole pairs travel toward the n-type semiconductor layer 215 and the p-type semiconductor layer 211, respectively, but in the APD shown in FIG. The difference is the energy step difference △Ec (for electrons, ΔEc for holes) for each step-back structure.
When Ev) becomes larger than the ionization energy, the electrons are ionized, generating electron-hole pairs and producing a multiplication effect. Of course, since each step-back structure layer has a similar effect, multiplication occurs by 2'' for the number n of layers.For example, ideally by setting ΔEc)ΔE y b50, the hole Since the ionization rate of is suppressed to be much lower than the ionization rate of electrons, the noise is lower than that of the above-mentioned APD.

[Problems to be Solved by the Invention] However, in order to put the APD described using FIGS. 43(a) to 43(c) into practical use, there were several technical issues to be solved.

Technical issues regarding the device's performance include (11) Since human radiation is absorbed by the p-type semiconductor layer and the multiplication layer, the multiplication factor changes depending on the incident wavelength of the light, making it unsuitable as a reading device.

(2) Since the forbidden band width of the light absorption layer and the multiplication layer is small, the dark current during operation is high and the noise is large.

(3) Since it is intended for optical communication, materials are limited, and the light that can be used is approximately 800 to 1600 nm, and cannot be used for light of other wavelengths such as visible light.

Technical issues in device fabrication include: (1) In order to create a step-back structure using a compound semiconductor, it is difficult to modulate the composition, and ΔEc.

There is a limit to the size of ΔEv, and there is a limit to noise reduction.

(2) Since the material is a compound semiconductor belonging to the III-V, II-TV group, etc., it has problems as an industrial material, such as material toxicity and cost.

(3) Methods for forming compound semiconductors have problems such as the need for ultra-high vacuum, the need to form films at high temperatures (approximately 500 to 650 degrees Celsius), and the difficulty of forming large areas. This is inappropriate as a manufacturing method for an element.

etc.

An object of the present invention is to solve the above-mentioned conventional technical problems, and to provide a photoelectric conversion device with a novel configuration that has excellent high-speed response, low noise and high sensitivity especially for visible light, and is easy to increase in area. The goal is to propose the following.

[Means for Solving the Problems] The photoelectric conversion device of the present invention includes a light absorption layer that has a forbidden band width Egl and absorbs light, and a minimum forbidden band width Eg2 that multiplies carriers generated by absorbing light. , a multiplication layer formed by laminating one or more layers of a step-back structure with a maximum forbidden band width Eg3 that continuously changes is sandwiched between charge injection blocking layers. shall be.

[Function] Hereinafter, the structure of the photoelectric conversion device and the structure of the energy band of the present invention will be explained using FIGS. 1(a) to (c), and the function of the present invention will be explained.

FIG. 1(a) is a schematic cross-sectional structural diagram showing the structure of the photoelectric conversion device of the present invention, in which an independent light absorption layer 310, a plurality of step-back structure layers 301, 303, which serve as multiplication layers,
305, 307, and 309 serve as charge injection blocking layers.
sandwiched between a p-type semiconductor layer 311 and an n-type semiconductor layer 315,
The type semiconductor device 11 and the electrode 313, the n-type semiconductor layer 315 and the electrode 314 are electrically connected, and the glass substrate 3
16. Note that the p-type semiconductor layer 311 serving as the charge injection blocking layer may be a metal that forms a Schottky junction with an adjacent semiconductor layer, which can expect the same effect as a matter of course. Further, although the case where the step-back structure layer has five layers is shown, the present invention is not limited to this, and may be one layer or two or more layers.

FIG. 1(b) is a schematic diagram of the energy band of the photoelectric conversion device when no bias is applied.

FIG. 1(c) is a schematic diagram of the energy band during reverse bias of the photoelectric conversion device.

The operating principle of the multiplication mechanism is similar to the conventional example proposed by Capasso et al., but the photoelectric conversion device of the present invention has the following effects.

(1) Step back layer 30 with independent light absorption layer 310
1 to 309 and a p-type semiconductor layer 311 which is a charge injection blocking layer.
Since it is sandwiched between the layers, light penetration into the multiplication layer is reduced, and there is little variation in the multiplication factor due to light penetration into the multiplication layer.

(2) Since the multiplication layer is composed of a step-back structure layer with a large ΔEc (ΔEv is large during electron multiplication and hole multiplication), low noise and a sufficient multiplication factor can be obtained.

(3) A non-single-crystal material is desirable as a constituent material of an element to which the present invention is applied. Here, the non-single crystal material is a polycrystalline material or an amorphous material, and the amorphous material is
The so-called microcrystalline structure is also included in this scope.

Specifically, amorphous silicon compensated with hydrogen and/or halogen elements (hereinafter referred to as a-Si (H,X))
, amorphous silicon germanium (hereinafter referred to as a-3iGe (
H,X)), amorphous silicon carbide (hereinafter referred to as a)
-3iC (H,
20] [311] It also includes amorphous silicon having crystallinity that has a peak specified by each Miller index.

In this way, since the constituent material of the element is a non-single crystal material, it can be processed at low temperatures (for example, 200 to 30
0°C) and can be easily fabricated on a large-area substrate, and the forbidden band width can be easily controlled (composition modulation, etc.), making it relatively easy to create a multiplication layer with a step-back structure, as well as atomic resistance caused by heat, etc. Thermal diffusion etc. are suppressed, a relatively reliable step-back structure can be formed, and problems in laminating multiple layers are reduced.

Further, in particular, the charge injection blocking layer can be made of a material with a relatively wide bandgap and amorphous silicon having high crystallinity with a high doping effect, so that dark current is reduced.

(4) Hydrogenated amorphous silicon or the like has a large light absorption coefficient, so the thickness of the light absorption layer can be reduced.

(5) Since the degree of freedom in the forbidden band width of the light absorption layer is increased for the same reason as in (3) above, a photoelectric conversion element with high sensitivity to incident light of various wavelengths can be constructed. Especially light absorption layer 3
By setting the forbidden band width Egl of 10 to a forbidden band width corresponding to visible light, high sensitivity to visible light can be achieved.

[Example] Hereinafter, an example of the present invention will be described in detail using the drawings.

(Example 1) Hereinafter, a first example of the present invention will be described using FIG. 2 and FIGS. 3(a) and (b).

FIG. 2 is a schematic vertical cross-sectional structural diagram showing a first embodiment of the photoelectric conversion device of the present invention.

In FIG. 2, 401 is a Cr electrode, 402 is an n-type a-3i+-x with a thickness of about 500 to prevent hole injection.
A charge injection blocking layer made of Gex:H, 403 is a-3t + -xGex:H~a for carrier multiplication.
-3it-, multiplication region with changed Cy:H composition, 4
04 has a thickness of approximately 2 to absorb light and generate carriers.
μm a-3i: A light absorption layer made of H, 405 has a thickness of about 100 μm to prevent electron injection, p-type a-3i:
The charge injection blocking layer 406 made of H is a transparent electrode mainly made of indium oxide.

The Cr electrode 401 and the transparent electrode 406 are made by EB evaporation, and the charge injection blocking layer 402. Multiplication region 403, light absorption layer 4
The amorphous layers of 04 and charge injection blocking layer 405 were formed by plasma CVD. The raw material gas for creating the amorphous layer is
Charge injection blocking layer 4゜2 is 5i) In, GeH4, PHx
, H2, multiplication region 403 is 5tli4°Ge)+4. C
H4, Hz, the light absorption layer 404 was made of SiH4, H2, and the charge injection blocking layer 405 was made of SiH4, B2H8, Ha.

The multiplication region 403 consists of three layers 411, 412, and 413 with a thickness of 200 layers in which the gas flow rates of CH4 and GeHa among the source gases are continuously changed.

The energy band structure of the photoelectric conversion device of the first embodiment shown in FIG. 2 is ideally assumed to be as shown in FIGS. 3(a) and 3(b).

FIG. 3(a) shows the energy band surface when the photoelectric conversion device of the first embodiment is in an unbiased state, and FIG. 3(b) shows the energy band surface when the photoelectric conversion device of the first embodiment is in a biased state to perform a carrier multiplication operation. is the energy band surface of

Figure 3 (a) and (b) are n-type a-3t+-xGe++
:The forbidden band width of the 8th layer 501 is Eg4, a-3t1-++G
ex:) I-a-3t+-yCy:H composition change layer 511
, 512, 513, the minimum forbidden band width of the multiplication region 502 is Eg2, the maximum forbidden band width of the multiplication region 502 is Eg3, the forbidden band width of the a-Si:8 layer 503 is Egl,
This shows that the forbidden band width of the p-type a-3i:H layer 504 is Ego.

In addition, in Figure 3(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen in Figure 3(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Of the composition change layers 511, 512, 513 created here, the layer giving the maximum forbidden band width Eg3 has a C composition ratio y of approximately 0.
, 4 a-3i+-yCy: H, and Eg3 was about 2.3 eV.

Also, a-3i+ -xGex: Ge of 8 layers 501
The composition ratio X is approximately 0.6, and the forbidden band width Eg4 is approximately 1.3.
It was eV. Among the composition change layers 511, 512, and 513, the layer giving the minimum forbidden band width Eg2 is also a-Si, □G
ex: 8 layers, and Eg2 was also about 1.3 eV. The forbidden band widths Egl and Ego of a-Si:H/il of 503°504 were both about 1.8 eV.

Further, the light absorption coefficient of the light absorption layer 503 is about I x 10 'cm-' or more for light with a wavelength of 400 nm, and the light absorption coefficient of the light absorption layer 503 is approximately I
It is about 5 x 10 "cm-' or more for light of 0 nm, and visible light can be sufficiently absorbed.

The multiplication factor of this device was about 10 times or more when a bias of 10V was applied.

Furthermore, for light with a wavelength of 700 nm or less, there was no change in the multiplication factor even when the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 502, and was fast.

In this example, there were three composition-change layers in the multiplication region, but this is just an example, and the number of layers can be any number (it can be determined according to the desired multiplication factor). .

Furthermore, in this example, a structure in which the stepback changes steeply is assumed as an ideal energy band diagram, but if the range is within the mean free path of the electron, the stepback becomes gradual. The same effect can be obtained. Furthermore, even if the stepback is gentle, it is still within the range that can work.

Furthermore, although the thickness of the composition change layer is approximately 200 in this embodiment, it may be within a range that allows carriers to travel without recombining. However, the thinner the material, the lower the applied bias (
It is desirable because it can be done.

Further, in this embodiment, the thickness of the light absorption layer is approximately 2 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

In addition, SiH is used as the raw material gas for the amorphous layer in this embodiment.

BJa, PH3, CH<, GeH4 was used, but SiH
Instead of +('), 5iFn, 512H8, 5i2F
a, 513H8゜5iHiF, 5izFi,...
chain silane compounds such as, or 5i5H+o+5iaH
Cyclic silane compounds such as +a, 514Hs, and ``'' can be used, and instead of B2H6, they contain group II atoms such as B (boron), A42 (aluminum), In (indium), and Tn (thallium). Gases can be used, and instead of PH, P (phosphorus), As (arsenic), Sb
(A gas containing Group 1 V atoms such as antimony L Bi (bismuth) can be used, and instead of CH, CH, F2.C2H8, C, H4,
Carbon compounds such as C, H2°Si(CH3)4.3iH(C:H3)s, N2. N.H.

Nitrogen compounds such as H2NNH2, HN, NH4N, F, N, F4N, Oa.

CO□, No, NO2, N20.0. ,N2O3,N2
Oxygen compounds such as O4 and No3 can be used, and instead of GeH4, germanium compounds such as GeF4 and tin compounds such as SnH4 can be used.

Furthermore, the composition ratio of the composition change layer is set to O to reduce localized levels.
A range of from about 0.6 is preferred.

In addition to the plasma CVD method, E
CR plasma method etc. are also useful.

Furthermore, in this example, an amorphous layer was used as the semiconductor layer, but
Non-single crystals such as polycrystals may also be used.

In this example, light is incident from the p-layer side of the charge injection blocking layer and is trapped in electrons to cause a multiplication operation. However, by replacing the p-layer and n-layer of the charge injection blocking layer, the valence electrons in the multiplication region are A step-back structure may be formed on the band side, and light may be incident from the n-layer side of the charge injection blocking layer to cause a multiplication operation by holes.

The forbidden band width Eg2 of the light absorption layer is approximately 1.8 eV here, but it is also possible to control Eg2 by changing the H2 gas flow rate so as to obtain desired spectral sensitivity characteristics.

In addition, the forbidden band width and doping amount of the charge injection blocking layer (both p and n) may be adjusted so that injection of minority carriers from the electrode can be suppressed and the mobility of majority carriers is not hindered. .

(Example 2) Hereinafter, a second example of the present invention will be described using FIG. 4.

FIGS. 4(a) and 4(b) are ideally assumed energy band structure diagrams of the second embodiment of the present invention.

FIG. 4(a) shows the energy output when the photoelectric conversion device of the second embodiment is in a non-biased state, and FIG. 4(b) shows the energy output when the photoelectric conversion device of the second embodiment is in a biased state for carrier multiplication. It is an energy empire.

In FIG. 4(a), 601 is an n-type a-Si + -, Cy:H layer with a forbidden band width Eg4', and 604 is a p-type a-3it-yCy:H layer with a forbidden band width Egg'. It is the same as FIG. 3(a) except that it is a layer, and a-3i
+ -xGex:H~a-3i+-yCy:H Multiplication region 6 consisting of three layers of composition change layers 611, 612, 613
02's minimum forbidden band width is Eg2, maximum forbidden band width is Eg3,
a-3i: Indicates that the forbidden band width of the 8-layer 603 is Egl.

The multiplication factor of this device was about 10 times or more when IOV bias was applied.

Further, even when the wavelength was changed for light having a wavelength of 700 nm or less, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was as low as about 0.1 nA/cm 2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin-type photoelectric conversion device without the multiplication layer 602, and was fast.

(Example 3) This example is an example in which the photoelectric conversion device shown in Example 1 is laminated on the scanning circuit and readout circuit that the present inventors have already proposed in Japanese Patent Application Laid-Open No. 63-278269. be.

FIG. 5(a) is a schematic cross-sectional view of the vicinity of the light receiving section of the embodiment of the present invention, FIG. 5(b) is an equivalent circuit diagram of one pixel, and FIG.
) is the other circuit and block circuit diagram of the entire device.

In FIG. 5(a), an n- layer 70 that becomes a collector region is epitaxially grown on an n-type silicon substrate 701.
2 is formed, in which a p base region 703 and an n base region 703 are formed.
An emitter region 704 is formed to constitute a bipolar transistor.

The p base region 703 is separated from adjacent pixels, and there is an oxide film 7 between the horizontally adjacent p base regions.
A gate electrode 706 is formed with '05 in between. Therefore, a p-channel MOS transistor is constructed by using the adjacent p base regions 703 as source and drain regions, respectively. Gate electrode 706 is p base region 703
It also works as a capacitor to control the potential of

Furthermore, after forming the insulating layer 707, the emitter electrode 70
8, and a base electrode 708'.

After that, an insulating layer 709 is formed, followed by an electrode 711, and each pixel is separated. Electrode 711 is electrode 708'
is electrically connected to.

Furthermore, eight layers 712 of n-type a-3i+-xGex are formed and separated for each pixel.

Subsequently, a-Si+-xGex:H~a-3it-,C
Composition change layers 721, 722, and 723 of y:H are formed to constitute a multiplication region 713. Next, light absorption layer a-3L+H
A transparent electrode 71 for forming a layer 714, forming a p-type a-3i:H layer 715, and applying a bias voltage to the sensor.
form 6.

Further, a collector electrode 717 is ohmically connected to the back surface of the substrate 701.

Therefore, the other circuit for one pixel is a bipolar transistor 7 made of crystalline silicon, as shown in FIG. 5(b).
A p-channel MOS transistor 732, a capacitor 733, and a photoelectric conversion device 734 similar to that in Example 1 are connected to the base of the p-channel MOS transistor 31, and a terminal 735 for applying a potential to the base, a p-channel MOS transistor 732, and a capacitor 733 are connected to the base of the p-channel MOS transistor 31. a terminal 736 for driving;
It is represented by a sensor electrode 737, an emitter electrode 738, and a collector electrode 739.

FIG. 5(c) is a circuit configuration diagram in which the pixel self cells 40 shown in FIGS. 5(a) and 5(b) are arranged in a 3×3 two-dimensional matrix.

In the figure, a collector electrode 741 of one pixel cell 40
are provided in all pixels, and sensor electrodes 742 are also provided in all pixels. Furthermore, the gate electrodes and capacitor electrodes of the PMO3 transistors are connected to drive wiring lines 743, 743'743'' for each row, and are connected to a vertical shift register (V, S, R) 744.

Further, the emitter electrodes are connected to vertical wirings 746, 746', 746'' for signal readout for each column.The vertical wirings 746, 746', 746'' are connected to switches 747,747' for resetting the charges of the vertical wirings, respectively. 7
47'' and readout switches 750°, 750', and 750''. Reset switch 747, 747',
The gate electrodes of 747'' are commonly connected to a terminal 748 for applying a vertical line reset pulse, and the source electrodes are commonly connected to a terminal 748 for applying a vertical line reset voltage.
9 are commonly connected. Read switch 750750
', 750'' gate electrodes are wires 751, 75, respectively.
1', 751'' to the horizontal shift register (H, S,
R) 752, and its drain electrode is connected to an output amplifier 757 via a horizontal readout wiring 753. The horizontal readout line 753 is connected to a switch 754 for resetting the charge of the horizontal readout line.

The reset switch 754 is connected to a terminal 755 for applying a horizontal wiring reset pulse and a terminal 756 for applying a horizontal wiring reset voltage.

Finally, the output of amplifier 757 is taken out from terminal 758.

The operation will be briefly explained below using FIGS. 5(a) to 5(c).

The incident light is absorbed by the light absorption layer 714 in FIG. 5(a), and the generated carriers are multiplied in the multiplication region 713.
It is accumulated within the base region 703.

When the drive pulse output from the vertical shift register in FIG. 5(c) appears on the drive wiring 743, the base potential increases via the capacitor, and signal charges corresponding to the amount of light are transferred from the pixels in the first row to the vertical wiring 746743. ', 746'', respectively.

Next, a scan pulse of 75 is sent from the horizontal shift register 752.
1,751', 751'', the switches 750, 750', 750'' are sequentially turned on and off, and the signal is taken out to the output terminal 758 through the amplifier 757. At this time, the set switch 754 is turned on while the switches 750 750' and 750- are turned on in order, and the residual charge on the horizontal wiring 753 is removed.

Next, vertical line reset switches 747°747', 7
47'' is in the ON state, and the vertical wiring 746, 746',
The residual charges of 746~ are removed. Then, when a negative direction pulse is applied to the drive wiring 743 from the vertical shift register 744, the PMO3 transistor of each pixel in the first row is turned off.
The N state is reached, the base residual charge of each pixel is removed, and the pixel is initialized.

Next, a drive pulse output from the vertical shift register 744 appears on the drive wiring 743', and the signal charges of the pixels in the second row are similarly taken out.

Next, the signal charges of the pixels in the third row are extracted in the same manner.

The device operates by repeating the above operations.

Note that in the embodiments described above, examples of circuits according to the inventions of the present inventors have been shown, but a general photoelectric conversion device may also be used.

Hereinafter, a case will be described in which the photoelectric conversion device according to the present invention is used in a photoelectric conversion device having a numerical configuration.

FIG. 6 is a block diagram showing a configuration when the present invention is applied to a photoelectric conversion device having a general configuration.

In the figure, reference numeral 801 indicates a plurality of photoelectric conversion units according to the present invention, for example, Example 1. The photoelectric conversion device of the present invention shown in Example 2 is used. The photoelectric conversion section 801 is connected to a signal output section 805. In the signal output section 805, 8
02 is an accumulation means for signal charges generated by the photoelectric conversion unit 801; 803 is a scanning means for scanning the signal charges; 80
4 amplifies the signal charge transferred by the scanning means 803.
This is a reading means consisting of a noise compensation circuit and the like.

Note that the accumulating means 802 is necessary when performing an accumulating operation, but may be omitted.

(Example 4) This example uses a light absorption layer that absorbs light and has a forbidden band width Egl, a minimum forbidden band width Eg2 and a maximum forbidden band width Eg3 that multiply carriers generated by absorbing light. A multiplication layer formed by laminating one or more layers having a step-back structure in which the forbidden band width changes continuously is laminated so as to be sandwiched between charge injection blocking layers, and the light absorption The forbidden band width Egl of the layer and the maximum forbidden band width Eg of the multiplication layer
3 are approximately equal.

In this example, by making the forbidden band width Egl of the light absorption layer and the maximum forbidden band width Eg3 of the multiplication layer substantially equal, band mismatching between the light absorption layer and the multiplication layer and various problems caused thereby are avoided. This eliminates the problem of high-speed response and prevents a decrease in high-speed response caused by obstacles to carrier mobility in the light-absorbing layer due to the generation of interfacial order, etc., achieving high-speed response and achieving high-speed response similar to that of a photodiode without a multiplication layer. By setting the forbidden band width Egl of the light absorption layer to a forbidden band width corresponding to ultraviolet light at the same time as responsiveness, high sensitivity to ultraviolet light can be achieved. Furthermore, in this embodiment, by selecting the number of calendars in the step-back structure, an amplification factor of 2 or more can be obtained, and low noise can be achieved.

Furthermore, the components of the photoelectric conversion device of this embodiment are at least S
If it is made of a single crystal containing i atoms, it becomes possible to easily control the forbidden band width and perform low-temperature stacking, thereby solving various problems caused by stacking.

A fourth embodiment of the present invention will be described below with reference to FIG. 7 and FIGS. 8(a) and (b).

Note that this example is similar to Example 1 except that the forbidden band width Egl of the light absorption layer and the maximum forbidden band width Eg3 of the multiplication layer are made approximately equal, and since the description will overlap, the explanation will be omitted. Parts are omitted.

FIG. 7 is a schematic vertical cross-sectional structural diagram showing a fourth embodiment of the photoelectric conversion device of the present invention.

In FIG. 7, 901 is a Cr electrode, and 902 is an n-type a-3it-x with a thickness of approximately 500 mm to prevent hole injection.
A charge injection blocking layer made of Gex:H, 903 is a-3i +□Gex:H~a-Si for carrier multiplication
A multiplication region in which the composition of t-yC,:H is changed; 904 is a light absorption layer made of a-St+-yCy:H with a thickness of approximately 1 gm for absorbing light and generating carriers; 905;
is about 100 thick p-type a- to prevent electron injection
3L+-yCy: Charge injection blocking layer made of H, 906
is a transparent electrode mainly made of indium oxide.

A Cr electrode 901 and a transparent electrode 906 are formed by EB deposition, and a charge injection blocking layer 902. Multiplication region 903, light absorption layer 9
The amorphous layers of 04 and charge injection blocking layer 905 were formed by plasma CVD. The raw material gas for creating the amorphous layer is
The charge injection blocking layer 902 is made of SiH4, GeH4, PHs,
Hz, the multiplication region 903 is SiH4°GeHa, CH4,
H2, the light absorption layer 904 was made of SiH4, CH+, H2, and the charge injection blocking layer 905 was made of SiH+, CH4, BJs, Hz.

The multiplication region 903 consists of three layers 911, 912, and 913 with a thickness of 200 layers in which the gas flow rates of CH4 and GeH4 among the source gases are continuously changed.

The energy band structure of the photoelectric conversion device of the fourth embodiment shown in FIG. 7 is ideally assumed to be as shown in FIGS. 8(a) and 8(b).

FIG. 8(a) shows the energy band when the photoelectric conversion device of the fourth embodiment is in an unbiased state, and FIG. 8(b) shows the energy band when the photoelectric conversion device of the fourth embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

Figures 8(a) and 8(b) show n-type a-3i+-xGex:
The forbidden band width of 8 layers 1001 is Eg4, a-3Ll-xGe
x:H-a-3ll-yCy:H composition change layer 1011,
Multiplication region 100 consisting of three layers of 1012 and 1013
The minimum forbidden band width of 2 is Eg2, the maximum forbidden band width of the multiplication region 1002 is Eg3, a-3i+ -yCy: H layer 100
The forbidden band width of the p-type a-3it-yCy:H layer 1004 is EgO.

In addition, in Figure 8(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen in Figure 8(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

a-3it-yCy created here: H layer 1004
The C composition ratio y of °1003 is approximately 0.4, and the forbidden band width E
Both gl and Ego were about 2.3 eV. Among the composition change layers 1011, 1012, and 1013, the layer providing the maximum forbidden band width Eg3 <> a-3x+-yCy: H, and Eg3 was also about 2.3 eV.

Also, a-3it-xGex: Ge of HN1001
The C composition ratio is approximately 0.6, and the forbidden band width Eg4 is approximately 1.3.
It was eV. Composition change layers 1011, 1012, 101
The layer giving the minimum forbidden band width Eg2 of 3 is also a-3i+
-xGex: 8 layers, and Eg2 was also about 1.3 eV.

Furthermore, the light absorption coefficient of the light absorption layer 1003 is at a wavelength of 540 nm.
About 4 x 10"cm-' for light of m, wavelength 35
It is about 3 x 10 'cm-' or more for light of 0 nm, and ultraviolet light can be sufficiently absorbed.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Further, for ultraviolet light having a wavelength of 400 nm or less, there was no change in the multiplication factor even if the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 1002, and was fast.

In this embodiment, the thickness of the light absorption layer is approximately 1 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 2.3 eV here, but it is also possible to control Eg2 by changing the C composition ratio so as to obtain desired spectral sensitivity characteristics.

(Embodiment 5) Hereinafter, a fifth embodiment of the present invention will be described using FIG. 9.

FIGS. 9(a) and 9(b) are ideally assumed energy band structure diagrams of the fifth embodiment of the present invention.

FIG. 9(a) shows the energy band when the photoelectric conversion device of the fifth embodiment is in an unbiased state, and FIG. 9(b) shows the energy band when the photoelectric conversion device of the fifth embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

In Fig. 9 (a>, 1101 is n of the forbidden band width Eg4'
Type a-3i, -, C,: Same as Fig. 8(a) except that it has 8 layers, and a-31+-hGex:
H-a-3L+-yCy: H composition change layer 1111, 11
The minimum forbidden band width of the multiplication region 1102 consisting of three layers of 12.1113 is Eg2, the maximum forbidden band width is Eg3, a-S1+
-yCy: The forbidden band width of the H layer 1103 is Egl, p-type a-
3il-11cy: indicates that the forbidden band width of the 8-layer 1104 is Eg○.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Further, even when the wavelength was changed for ultraviolet light having a wavelength of 400 nm or less, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was as low as about 0.1 nA/cm 2 or less when a bias of 10 V was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 1102, and was fast.

(Example 6) Similar to Example 3, in this example, the photoelectric conversion device shown in Example 4 was already developed by the present inventors in Japanese Patent Application Laid-Open No. 63278269.
This is a structure laminated on the scanning circuit and readout circuit proposed in the publication.

Note that the configuration of the photoelectric conversion device is such that the light absorption layer a-3i:H layer 714 in FIG. 5(a) is replaced by the light absorption layer a-3i1.
-, Cy: 8 layers, p-type a-3i: H715 is p-type a
The structure is the same except that the -3i+-yCy+H layer is used, and the operation thereof is also the same, so a description thereof will be omitted.

Further, in the embodiment described above, an example of the circuit according to the invention of the present inventors has been shown, but as in the third embodiment, a photoelectric conversion device for boat fishing may also be used.

(Example 7) This example uses a light absorption layer that absorbs light and has a forbidden band width Egl, a minimum forbidden band width Eg2 and a maximum forbidden band width Eg3 that multiply carriers generated by absorbing light. A multiplication layer formed by laminating one or more layers having a step-back structure in which the forbidden band width changes continuously is laminated so as to be sandwiched between charge injection blocking layers, and the light absorption The forbidden band width Egl of the layer and the minimum forbidden band width Eg of the multiplication layer
2 is approximately equal.

In this example, by making the forbidden band width Egl of the light absorption layer and the minimum forbidden band width Eg2 of the multiplication layer substantially equal, band mismatching between the light absorption layer and the multiplication layer and various problems caused thereby are avoided. This eliminates the problem of high-speed response and prevents a decrease in high-speed response caused by obstacles to carrier mobility in the light-absorbing layer due to the generation of interfacial order, etc., achieving high-speed response and achieving high-speed response similar to that of a photodiode without a multiplication layer. By making the forbidden band width Egl of the light absorption layer particularly compatible with infrared light at the same time as responsiveness can be obtained, high sensitivity to infrared light can be provided.

Furthermore, in this embodiment, the amount of light incident on the multiplication layer is reduced, making it possible to reduce variations in the multiplication factor due to light incident on the multiplication layer.

Furthermore, in this embodiment, by selecting the number of calendars in the step-back structure, an amplification factor of 2 or more can be obtained, and low noise can be achieved.

Furthermore, if the components of the photoelectric conversion device of this embodiment are constructed from a single crystal containing at least Si atoms, easy control of the forbidden band width and low-temperature stacking become possible, and various problems caused by stacking can be solved. I can do it.

Hereinafter, a seventh embodiment of the present invention will be described using FIG. 10 and FIGS. 11(a) and (b).

This example is the same as Example 1 except that the forbidden band width Egl of the light absorption layer and the minimum forbidden band width Eg2 of the multiplication layer are made approximately equal, and since the description will overlap, the explanation will be omitted. Parts are omitted.

FIG. 10 is a schematic vertical cross-sectional structural diagram showing a seventh embodiment of the photoelectric conversion device of the present invention.

In Fig. 10, 1201 is a Cr electrode, 1202 is an n-type a-Si with a thickness of about 500 mm to prevent hole injection.
l-hGex: charge injection blocking layer made of H, 12
03 is a-3i+-xGex for carrier multiplication
:H~a-sll□C, +H multiplication region with different composition; 1204 is a light absorption layer made of a-3it□GeX:H with a thickness of about 1 μm for absorbing light and generating carriers; 1205 has a thickness of approximately 100 mm to prevent electron injection.
The charge injection blocking layer 1206 made of human p-type a-3x+-xGex:H is a transparent electrode mainly made of indium oxide.

The Cr electrode 1201 and the transparent electrode 1206 are created by EB deposition, and the charge injection blocking layer 1202. The amorphous layers of the multiplication region 1203, the light absorption layer 1204, and the charge injection blocking layer 1205 were formed by plasma CVD. The raw material gas for forming the amorphous layer is such that the charge injection blocking layer 1202 is made of SiH4, Ge.
H4, PHi, Hz, multiplication region 1203 is 5IH4Ge
H4, CHn, Hz, the light absorption layer 1204 is SiH4,
GeH4, H2, charge injection blocking layer 1205 is SiH4,
GeH4, B2Hs, and Hz were used.

The multiplication region 1203 consists of three layers 1211, 1212, and 1213 with a thickness of 200 layers in which the gas flow rates of CH4 and GeH4 among the source gases are continuously changed.

The energy band structure of the photoelectric conversion device of the seventh embodiment shown in FIG. 10 is ideally as shown in FIG. 11(a).

It is determined that it is as shown in (b).

FIG. 11(a) shows the energy band when the photoelectric conversion device of the seventh embodiment is in an unbiased state, and FIG. 11(b) shows the energy band when the photoelectric conversion device of the seventh embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

Figure 11 (a) and (b) are n-type a-3i + -xG
ex: Forbidden band width of 8 layers 1301 is Eg4, a-3i+
-xGex: H ~ a-3i + -yCy: H The minimum forbidden band width of the multiplication region 1302 consisting of three layers 1311, 1312, and 1313 is Eg2, multiplication region 1
The maximum forbidden band width of 302 is Eg3, a-3i + -xG
ex: Forbidden band width of 8 layers 1303 is Egl, p type a-3
i+-xGex: indicates that the forbidden band width of the 8th layer 1304 is Eg○.

In addition, in Fig. 11(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen from Fig. 11(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Composition change layer 131.1.1312131 created here
3, the layer giving the maximum forbidden band width Eg3 has a C composition ratio y
is about 04 a-Sit-yCy: H, and Eg3
was about 2.3 eV.

Also, a-3i+-xGex: 8 layers 1301, 13
The Ge composition ratio X of 03°1304 is approximately 0.6, and the forbidden band width Eg4. Egl and Eg○ were approximately 1.3 eV.

Among the composition change layers 1311, 1312, and 1313, the layer that provides the minimum forbidden band width Eg2 is also a-Sit-xGex:
It was an H layer, and Eg2 was also about 1.3 eV.

Furthermore, the light absorption coefficient of the light absorption layer 1303 is at a wavelength of 800 nm.
About I x 105 cm-' for light of m, wavelength 11
It is about 2×10′cm−′ or more for light of 000n, and infrared light is sufficiently absorbed.

The multiplication factor of this device was about 10 times or more when a bias of 10V was applied.

Further, for light having a wavelength of 11,000 nm or less, there was no change in the multiplication factor even if the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 10 nA/am2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 1302, and was fast.

In this embodiment, the thickness of the light absorption layer is approximately 1 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 1.3 eV here, but it is also possible to control Eg2 by changing the Ge composition ratio so as to obtain desired spectral sensitivity characteristics.

(Embodiment 8) Hereinafter, an eighth embodiment of the present invention will be described using FIG. 12.

FIGS. 12(a) and 12(b) are ideally determined energy band structure diagrams of the eighth embodiment of the present invention.

FIG. 12(a) shows the energy band when the photoelectric conversion device of the eighth embodiment is in an unbiased state, and FIG. 12(b) shows the energy band when the photoelectric conversion device of the eighth embodiment is in a biased state to perform carrier multiplication operation. The energy band is the same.

In FIG. 12(a), 1401 is n of the forbidden band width Eg4'
Type a-5i+-, Cy: 11th layer except that it is 8 layers.
Same as figure (a), a-3t+-xGex:H~a
-3it-yCy: The minimum forbidden band width of the multiplication region 1402 consisting of three layers of H composition change layers 1411, 1412° and 1413 is Eg2, the maximum forbidden band width is Eg3, a-3i + −
xGex: The forbidden band width of the H layer 1403 is Egl, p-type a
-S1+-++Ge+: Forbidden band width of 8 layers 1404 is E
g.

It shows that.

The multiplication factor of this device was approximately 10 times or more when a bias of 10 μ was applied.

Further, even when the wavelength was changed for infrared light having a wavelength of 11,000 nm or less, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was as low as about 10 nA/cm2 or less when a bias of 10 V was applied.

Furthermore, the photoresponse speed is higher than that of p i without the multiplication layer 1402.
It was equivalent to an n-type photoelectric conversion device and was faster.

(Example 9) Similar to Example 3, in this example, the photoelectric conversion device shown in Example 7 was already developed by the present inventors in Japanese Unexamined Patent Publication No. 63-27826.
This is an embodiment in which the present invention is stacked on the scanning circuit and readout circuit proposed in Publication No. 9. The configuration of the photoelectric conversion device is shown in Figure 5 (
a-3i as the light absorption layer of a): l (layer 714 is the light absorption layer as a-3i +-xGex: 8 layers, p-type a-3
i:H715 is p-type a-3i+-i+Gex: The structure is the same except that it has 8 layers, and its operation is also the same, so a description thereof will be omitted.

Further, in the embodiment described above, an example of a circuit according to the invention of the present inventors has been shown, but as in the third embodiment, a -numerical photoelectric conversion device may also be used.

(Example 10) This example includes a light absorption layer that absorbs light and has a forbidden band width Egl, a minimum forbidden band width Eg2 and a maximum forbidden band width Eg3 that multiply carriers generated by absorbing light. A multiplication layer formed by laminating one or more layers having a step-back structure in which the forbidden band width changes continuously is laminated so as to be sandwiched between charge injection blocking layers, and the light absorption The forbidden band width Egl of the layer is continuously changed from the side of one charge injection layer laminated on the light absorption layer, so that it becomes approximately equal to the maximum forbidden band width Eg3 of the multiplication layer on the multiplication layer side. This is what I did.

In this example, the forbidden band width Egl of the light absorption layer is changed continuously from the charge injection layer side stacked on the light absorption layer, and the maximum forbidden band width of the multiplication layer is changed on the multiplication layer side. By making Eg3 approximately equal to Eg3, band mismatch between the light absorption layer and the multiplication layer and various problems caused by it can be resolved, and interference with carrier mobility in the light absorption layer due to the generation of interface order can be avoided. This prevents the deterioration in high-speed response that occurs, achieves high-speed response, and provides high-speed response similar to that of a photodiode without a multiplication layer. At the same time, the forbidden band width E of the light absorption layer is
High sensitivity from visible light to ultraviolet light can be achieved by specifically setting gl to a forbidden band width corresponding to visible light to ultraviolet light.

Further, in this embodiment, by selecting the number of layers of the step-back structure, an amplification factor of 2 or more can be obtained and low noise can be achieved.

Further, the components of the photoelectric conversion device of this embodiment are at least S
If it is made of a single crystal containing i atoms, it becomes possible to easily control the forbidden band width and perform low-temperature stacking, thereby solving various problems caused by stacking.

Hereinafter, a tenth embodiment of the present invention will be described using FIG. 13 and FIGS. 14(a) and (b).

In addition, in this example, the forbidden band width Egl of the light absorption layer is changed continuously from the side of one charge injection layer laminated on the light absorption layer, and the maximum forbidden width of the multiplication layer is changed on the side of the multiplication layer. The second embodiment is the same as the first embodiment except that the band width Eg3 is made substantially equal to the band width Eg3, and since the description will be repeated, some of the description will be omitted.

FIG. 13 is a schematic vertical cross-sectional structural diagram showing a tenth embodiment of the photoelectric conversion device of the present invention.

In FIG. 13, 1501 is a Cr electrode, 1502 is an n-type a-3i with a thickness of approximately 500 mm to prevent hole injection.
+ -XGeX: Charge injection blocking layer made of H, 15
03 is a-3L+□Ge for carrier multiplication:
H~a-3t+ -yCy: A multiplication region with a changed composition of H, 1504 is a-3i:H-a-3it-yC with a thickness of about 2 trm for absorbing light and generating carriers.
y: A light absorption layer with a changed H composition, 1505 is a p-type a-3i with a thickness of about 100 to prevent electron injection.
:H charge injection blocking layer 1506 is a transparent electrode mainly made of indium oxide.

The Cr electrode 1501 and the transparent electrode 1506 are made by EB evaporation, and the charge injection blocking layers 15, 02. Multiplication area 1503,
The amorphous layers of the light absorption layer 1504 and the charge injection blocking layer 1505 were formed by plasma CVD. The raw material gas for forming the amorphous layer is 5 it (4,
GeHa, PH3, H2, multiplication region 1503 is 5iH
t, GeH4,CH<, Hz, the light absorption layer 1504 is SiH4,CH4,H2, and the charge injection blocking layer 1505 is S
iH4°B, H,, H2 was used.

The multiplication region 1503 is composed of three layers: 200-layer composition change layers 1511, 1512, and 1513, in which the gas flow rates of CH4 and GeHa among the source gases are continuously changed. The light absorption layer 1504 is formed by continuously changing the gas flow rate of CI (4) among the source gases.

Ideally, the energy band structure of the photoelectric conversion device of the tenth embodiment shown in FIG. 13 is as shown in FIGS. 14(a) and (b).
).

FIG. 14(a) shows energy storage when the photoelectric conversion device of the 10th embodiment is in an unbiased state, and FIG. 14(b) shows energy storage when a bias is applied to perform a carrier multiplication operation. It is an energy band.

Figures 14(a) and 14(b) show n! 8! a-3it-x
Gex: The forbidden band width of the H layer 1601 is E'g4, a-3i
+ -xGex:H~a-3il-1'clJ:
The minimum forbidden band width of the multiplication region 1602 consisting of three layers of H composition change layers 1611, 1612.16'13 is Eg2,
The maximum forbidden band width of the multiplication region 1602 is Eg3, a-3i:
H-a-5it-yCy: The minimum forbidden band width of the 8 layer 1603 is Egl, and the forbidden band width of the p-type a-3i+H layer 1604 is E
It shows that it is gO.

Also, in Figure 14(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen in Figure 14(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Composition change layers 1611, 16121613 created here
The layer giving the maximum forbidden band width Eg3 is a-Si+-yCy:H with a C composition ratio y of about 0.4, and Eg3
was about 2.3 eV. a-3i:1(-a-3i+
-yCy: The layer giving the maximum forbidden band width of 8 layers 1603 is also a
-3i+-y Cy20.

Also, a-3i+-xGex: Ge of 8 layers 1601
The composition ratio X is approximately 0.6, and the forbidden band width Eg4 is approximately 1.3.
It was eV. Composition change layer 1611, 1612.161
The layer giving the minimum forbidden band width Eg2 of 3 is also a-5it.
-++Gex: 8 layers, Eg2 is also about 1.3eV
Met. a-5i: H~a-5i+-yCy: 8 layers 1
The layer that provides the minimum forbidden band width Egl of 603 is a-3i:
H, and Egl was about 1.8 eV. p type a-3
i: The forbidden band width Ego of the 8-layer 1604 is also approximately 1.8 eV
Met.

Furthermore, the optical absorption coefficient of optical absorption N1603 is at a wavelength of 700 n
About 6 x 10"cm-' for light of m, wavelength 35
It is about 3 x 10 'cm-' or more for light of 0 nm, and absorption of visible to ultraviolet light is sufficiently performed.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Furthermore, there was no change in the multiplication factor even when the wavelength was changed for light from visible to ultraviolet wavelengths of 700 nm or less.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 1602, and was fast.

In this embodiment, the thickness of the light absorption layer was approximately 2 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 1.8 to 2.3 here.
eV, but Eg2 is controlled by changing the C composition ratio,
It is also possible to obtain desired spectral sensitivity characteristics.

(Embodiment 11) Hereinafter, the first embodiment of the present invention will be described using FIG. 15.

FIGS. 15(a) and 15(b) are ideally assumed energy band structure diagrams of the eleventh embodiment of the present invention.

FIG. 15(a) shows the energy band when the photoelectric conversion device of the eleventh embodiment is in an unbiased state, and FIG. 15(b) shows the energy band when the photoelectric conversion device of the eleventh embodiment is in a biased state to perform carrier multiplication operation. The energy band is the same.

In FIG. 15(a), 1701 is n of the forbidden band width Eg4'.
Type a-3it-, C, :H layer and light absorption layer 1
14(a) except that the a-3L:H composition region of 703 is wider than that in FIG. 14, and a-Sit
-xGex:H~a-3it-yCy:H composition change layer 1
Multiplication region 1 consisting of three layers of 711.1712.1713
The minimum forbidden band width of 702 is Eg2, and the maximum forbidden band width is Eg3.
, a-3i: H~a-3i+-, C,: The minimum forbidden band width of the H layer 1703 is Egl, p-type a-3i: 8 layers 1704
This shows that the forbidden band width of is EgO.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Furthermore, even when the wavelength was changed from the visible wavelength region of 700 nm or less to the ultraviolet wavelength region, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 1702, and was fast.

(Example 12) Similar to Example 3, in this example, the photoelectric conversion device shown in Example 10 was already developed by the present inventors in Japanese Unexamined Patent Publication No. 63-2782.
This structure is stacked on the scanning circuit and readout circuit proposed in Japanese Patent No. 69.

Note that the configuration of the photoelectric conversion device is as shown in FIG.
H=a-3it-yCy:The structure is the same except that the H composition change layer is used, and the operation thereof is also the same, so a description thereof will be omitted.

Further, in the embodiment described above, an example of a circuit according to the invention of the present inventors has been shown, but as in the third embodiment, a -numerical photoelectric conversion device may also be used.

(Example 13) This example includes a light absorption layer that absorbs light and has a forbidden band width Egl, a minimum forbidden band width Eg2 and a maximum forbidden band width Eg3 that multiply carriers generated by absorbing light. A multiplication layer formed by laminating one or more layers having a step-back structure in which the forbidden band width changes continuously is laminated so as to be sandwiched between charge injection blocking layers, and the light absorption The forbidden band width Egl of the layer is continuously changed from one charge injection layer side laminated on the light absorption layer, so that on the multiplication layer side it becomes approximately equal to the minimum forbidden band width Eg2 of the multiplication layer. This is what I did.

In this example, the forbidden band width Egl of the light absorption layer is changed continuously from the side of one charge injection layer laminated on the light absorption layer, and the minimum forbidden band width of the multiplication layer is changed on the multiplication layer side. By making Eg2 approximately equal to Eg2, band mismatch between the light absorption layer and the multiplication layer and various problems caused by it can be resolved, and interference with the mobility of carriers in the light absorption layer due to the generation of interface order can be avoided. This prevents the deterioration in high-speed response that occurs, achieves high-speed response, and provides high-speed response similar to that of a photodiode without a multiplication layer. At the same time, the forbidden band width E of the light absorption layer is
High sensitivity from infrared light to ultraviolet light can be achieved by setting gl to a forbidden band width that corresponds to in particular from infrared light to ultraviolet light.

Further, in this embodiment, by selecting the number of layers of the step-back structure, an amplification factor of 2 or more can be obtained and low noise can be achieved.

Further, the components of the photoelectric conversion device of this embodiment are at least S
If it is made of a single crystal containing i atoms, it becomes possible to easily control the forbidden band width and perform low-temperature stacking, thereby solving various problems caused by stacking.

A thirteenth embodiment of the present invention will be described below with reference to FIG. 16 and FIGS. 17(a) and (b).

In addition, in this example, the forbidden band width Egl of the light absorption layer is continuously changed from the side of one charge injection layer laminated on the light absorption layer, and the minimum forbidden width of the multiplication layer is changed on the side of the multiplication layer. It is the same as in Example 1 except that it is made to be approximately equal to the band width Eg2, and since the description will be repeated, a part of the description will be omitted.

FIG. 16 is a schematic vertical cross-sectional structural diagram showing a thirteenth embodiment of the photoelectric conversion device of the present invention.

In FIG. 16, 1801 is a Cr electrode, 1802 is an n-type a-3i with a thickness of approximately 500 mm to prevent hole injection.
+ -xGex: charge injection blocking layer made of H, 18
03 is a-3i + ~ xGe for carrier multiplication
x:H~a-Si+-yCy: Multiplication region with a changed composition of H, 1804 is a thickness of approximately ] μm for absorbing light and generating carriers a-3i:H~a-3i+-xG
ex: A light absorption layer with a changed composition of H, 1805 is a p-type a-3i with a thickness of about 100 to prevent electron injection:
The charge injection blocking layer 1806 made of H is a transparent electrode mainly made of indium oxide.

A Cr electrode 1801 and a transparent electrode 1806 are formed by EB evaporation, and a charge injection blocking layer 1802. The amorphous layers of the multiplication region 1803, the light absorption layer 1804, and the charge injection blocking layer 1805 were formed by plasma CVD. The raw material gas for forming the amorphous layer is 5L for the charge injection blocking layer 1802)14. G
eH4, PHi, H2, multiplication region 1803 is SiH4,
GeH4, CH4, H2, light absorption layer 1804 is SiH4
, GeH4, Hz, and the charge injection blocking layer 1805 is SiH4.
, B2H,, H2 were used.

The multiplication region 1803 consists of three layers 1811, 1812, and 1813 with a thickness of 200 layers in which the gas flow rates of CH and GeH4 among the source gases are continuously changed. The light absorption layer 1804 is formed by continuously changing the gas flow rate of GeH4 among the raw material gases.

The energy band structure of the photoelectric conversion device of the thirteenth embodiment shown in FIG. 16 is ideally as shown in FIGS. 17(a) and (b).
).

FIG. 17(a) shows the energy output when the photoelectric conversion device of the 13th embodiment is in a non-biased state, and FIG. 17(b) shows the energy output when the photoelectric conversion device of the 13th embodiment is in a biased state to perform carrier multiplication operation. It is an energy empire.

FIG. 17(a)(b) shows n-type a-3i+-xGex:
The forbidden band width of the 8th layer 1901 is Eg4, a-3it-xGe
x:H~a-3L-yCy:H composition change layer 1911,1
Multiplication region 1902 consisting of three layers of 912.1913
The minimum forbidden band width of Eg2, the maximum forbidden band width of the multiplication region 1902 is Eg3, a-3i:H~a-3i+-xGex+
The maximum forbidden band width of the H layer 1903 is Egl, p-type a-Si
: Indicates that the forbidden band width of the 8th layer 1904 is Ego.

Also, in Figure 17(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen in Figure 17(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Composition change layer 1911.19121913 created here
The layer giving the maximum forbidden band width Eg3 was a-3it-yCy:H with a C composition ratio y of about 04, and Eg3 was about 2.3 eV.

Also, a-Si, -xGet: 8 layers 19o1 Ge
The composition ratio X is approximately 0.6, and the forbidden band width Eg4 is approximately 1.3.
It was eV. Composition change layer 1911, 1912.191
The layer giving the minimum forbidden band width Eg2 of 3 is also a-31+
-xGe++: 8 layers, Eg2 is also about 1.3 eV
Met. a-3i: H-a-3i+-xGex: The layer giving the minimum forbidden band width of 8 layers 19o3 is also a-Si+-+
+Gex:H.

The layer giving the maximum forbidden band width Egl of the a-3i:Hza-3it-xGex:8 layer 19o3 was a-3i:H, and Egl was about 1.8 eV. The forbidden band width Ego of the p-type a-3i:8 layer 19o4 was also about 1.8 eV.

Furthermore, the light absorption coefficient of the light absorption layer 1903 is at a wavelength of 400 nm.
m light, approximately I x 105 cm-' or more, wavelength 1
It is about 2×10 cm −′ or more for 1000 nm of light, and infrared, visible, and ultraviolet light can be sufficiently absorbed.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Furthermore, for light with a wavelength of 11,000 nm or less, there was no change in the multiplication factor even if the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm 2 or less when a bias of 10 μm was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 1902, and was high.

Further, in this embodiment, the thickness of the light absorption layer is approximately 1 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 1.8 to 1.3 here.
eV, but control Eg2 by changing the Ge composition ratio,
It is also possible to obtain desired spectral sensitivity characteristics.

(Embodiment 14) Hereinafter, a 14th embodiment of the present invention will be described using FIG. 18.

FIGS. 18(a) and 18(b) are ideally assumed energy band structure diagrams of the fourteenth embodiment of the present invention.

FIG. 18(a) shows the energy band when the photoelectric conversion device of the 14th embodiment is in a non-biased state, and FIG. 18(b) shows the state in which a noise voltage is applied to perform a carrier multiplication operation. The energy band of the time is the same.

In FIG. 18(a), 2001 is n of the forbidden band width Eg4'.
It is the same as FIG. 17(a) except that it is a type a-5i + -yCy:H layer and the a-3i:H composition region of the light absorption layer 2003 is wider than that in FIG. 17, a-
The minimum forbidden band width of the multiplication region 2002 consisting of three layers of 3t+-xGex:H~a-3t+-, Cy:H composition change layer 20112012.2013 is Eg2, and the maximum forbidden band width is E
g3, a-3i:l(-a-3it-xGex:8 layer 2
The maximum forbidden band width of 003 is Egl, and the maximum forbidden band width of p-type a-3i:0 layer 2004 is Ego.

The multiplication factor of this device was about 10 times or more when a bias of 10V was applied.

Further, even when the wavelength was changed for light having a wavelength of 11,000 nm or less, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm 2 or less when a bias of 10 V was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 2002, and was fast.

(Example 15) Similar to Example 3, in this example, the photoelectric conversion device shown in Example 13 was already developed by the present inventors in Japanese Unexamined Patent Publication No. 63-2782.
This structure is stacked on the scanning circuit and readout circuit proposed in Japanese Patent No. 69. The configuration of the photoelectric conversion device is shown in Figure 5 (
It is the same except that the a-3i:H layer 714 that is the light absorption layer in a) is replaced with the a-3i:H-a-3it-xGex:H composition change layer that is the light absorption layer, and its operation is also the same. Therefore, the explanation will be omitted.

Further, in the embodiment described above, an example of a circuit according to the invention of the present inventors has been shown, but as in the third embodiment, a -numerical photoelectric conversion device may also be used.

(Example 16) This example consists of a light absorption layer that has a forbidden band width Egl and absorbs light, a metal electrode, a minimum forbidden band width Eg2 that multiplies carriers generated by absorbing light, and a maximum A multiplication layer formed by laminating one or more layers having a step-back structure in which the forbidden band width Eg3 continuously changes, and a multiplication layer laminated in sequence are stacked so as to be sandwiched between charge injection blocking layers. This is what I did.

In this example, a layer in which a light absorption layer, a metal electrode, and a multiplication layer are sequentially stacked is sandwiched between charge injection blocking layers, thereby eliminating band mismatch between the light absorption layer and multiplication layer, and It eliminates various problems caused by the formation of interfacial order, prevents the decline in high-speed response caused by obstacles to carrier mobility in the light absorption layer due to the generation of interfacial order, achieves high-speed response, and does not have a multiplication layer. At the same time, even if light incident from the light absorption layer side passes through the light absorption layer, it will not enter the multiplication layer because the metal electrode will block the light, and the light will not enter the multiplication layer. It is possible to eliminate fluctuations in the multiplication factor caused by Furthermore, in this embodiment, by setting the forbidden band width Egl of the light absorption layer to various sizes, it is possible to provide high sensitivity not only to visible light but also to light of various wavelengths.

Further, in this embodiment, by selecting the number of layers of the step-back structure, an amplification factor of 2 or more can be obtained and low noise can be achieved.

Furthermore, if the light absorption layer, multiplication layer, and charge injection blocking layer of the photoelectric conversion device of this example are made of a single crystal containing at least Si atoms, easy control of the forbidden band width and low-temperature lamination become possible. Various problems caused by this can be solved.

Hereinafter, a sixteenth embodiment of the present invention will be described using FIG. 19 and FIGS. 20(a) and (b).

Note that this example is the same as Example 1, except that a layer in which a light absorption layer, a metal electrode, and a multiplication layer are sequentially laminated is sandwiched between charge injection blocking layers, and the explanation will be repeated. Therefore, some explanations are omitted.

FIG. 19 is a schematic vertical cross-sectional structural diagram showing a 16th embodiment of the photoelectric conversion device of the present invention.

In FIG. 19, 2101 is a Cr electrode, 2102 is an n-type a-3i with a thickness of approximately 500 mm to prevent hole injection.
+-xGex:H charge injection blocking layer, 2103 is a-3i for carrier multiplication +□Gex:H~
a-3i+-yCy: A multiplication region with a different H composition, 2104 is a Cr electrode with a thickness of approximately 200 mm to prevent light from penetrating into the multiplication region, and 2105 is a Cr electrode that absorbs light and carries carriers. A-3i with a thickness of about 1 μm for generation:
A light absorption layer 2106 is made of H, a charge injection blocking layer made of p-type a-3i:H having a thickness of approximately 100 nm for blocking electron injection, and 2107 is a transparent electrode mainly made of indium oxide.

The Cr electrodes 2101 and 2104 and the transparent electrode 2107 are E
Created by B vapor deposition, charge injection blocking layer 2102, multiplication region 2
103. Light absorption layer 2105 and charge injection blocking layer 210
The amorphous layer of No. 6 was created by plasma CVD method. The source gas for forming the amorphous layer is such that the charge injection blocking layer 2102 is Si
H+, GeH, PH.

H2, multiplication region 2103 is SiH4, GeH4, CH4
, Hz, the light absorption layer 2105 was made of SiH4, H2, and the charge injection blocking layer 2106 was made of SiH4, BJs, Hz.

The multiplication region 2103 is made up of three layers: 200-layer composition change layers 2111, 2112, and 2113, in which the gas flow rates of CH4 and GeH4 among the source gases are continuously changed.

The energy band structure of the photoelectric conversion device of the 16th embodiment shown in FIG. 19 is ideally as shown in FIGS.
).

Figure 20(a) shows the energy output when the photoelectric conversion device of the 16th embodiment is in a non-biased state, and Figure 20(b) shows the energy output when the photoelectric conversion device of the 16th embodiment is in a biased state to perform carrier multiplication operation. It is an energy empire.

Figure 20 (a) and (b) are n-type a-3it-++Ge
x: H layer 2201 stratification 201Eg4, a-3t+-x
Gex:H~a-3i+-ycy:H composition change layer 221
Multiplication region 2 consisting of three layers of 1,2212.2213
The minimum forbidden band width of 202 is Eg2, the maximum forbidden band width of the multiplication region 2202 is Eg3, the forbidden band width of a-5i:H layer 2204 is Egl, and the forbidden band width of p-type a-3i:H layer 2205 is EgO. It shows that. 2203 corresponds to a Cr electrode.

Also, in Figure 20(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen in Figure 20(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Composition change layer 2211.22122213 created here
The layer giving the maximum forbidden band width Eg3 is a-3j with a C composition ratio y of about 0.4, + -yCy:H,
Eg3 was about 2.3 eV.

In addition, a-3it-xGex: Ge of the H layer 2201
The composition ratio X is approximately 0.6, and the forbidden band width Eg4 is approximately 1.3.
It was eV. Composition change layer 2211, 2212.221
The layer giving the minimum forbidden band width Eg2 of 3 is also a-3i+
-xGex: H layer, and Eg2 was also about 1.3 eV. 2204.2205 a-3i: The forbidden band widths Egl and Ego of the H layer were both about 1.8 eV.

Furthermore, the light absorption coefficient of the light absorption layer 2203 is at a wavelength of 400 nm.
About I x 105 cm-' for light of m, wavelength 70
It is about 5 X I O"cm-' or more for light of 0 nm, and visible light can be sufficiently absorbed.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Furthermore, for visible light with a wavelength of 700 nm or less, there was no change in the multiplication factor even if the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm'' or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin-type photoelectric conversion device without the multiplication layer 2202, and was fast.

Further, in this example, the thickness of the light absorption layer was approximately 1 μm, but it may be any thickness that provides the desired spectral sensitivity. This thickness is determined by the light absorption coefficient.

In addition, in this example, the electrode sandwiched between the light absorption layer and the multiplication layer was made of Cr, but this electrode material is used to transport electrons from the light absorption layer to the multiplication layer and from the multiplication layer to the light absorption layer. Any material may be used as long as it has a work function that allows holes to be smoothly transported to and has excellent light shielding properties. For example, the electrode materials include Mg, Aff, Ti, Mn, Fe, Cu.
, Zn, Mo, Ag, Cd, In, Sn, W, and alloys thereof.

The forbidden band width Eg2 of the light absorption layer is approximately 1.8 eV here, but it is also possible to control Eg2 by changing the H2 gas flow rate so as to obtain desired spectral sensitivity characteristics.

(Example 17) Hereinafter, a seventeenth example of the present invention will be described using FIG. 21.

FIGS. 21(a) and 21(b) are ideally assumed energy band structure diagrams of the seventeenth embodiment of the present invention.

FIG. 21(a) shows the energy consumption when the photoelectric conversion device of the 17th embodiment is in a non-biased state, and FIG. 21(b) shows the energy consumption when the photoelectric conversion device of the 17th embodiment is in a biased state for carrier multiplication. It is an energy empire.

In FIG. 21(a), 2301 is n of the forbidden band width Eg4'
It is the same as FIG. 20(a) except that it is a type a-Si+-yCy:8 layer and 2305 is a p-type a-3it-yC,:8 layer with a forbidden band width Ego'. a-S
i1-xGex:H~a-3x+-yC, :H The multiplication region 2302 consisting of three layers 2311, 2312, and 2313 has a minimum forbidden band width of Eg2 and a maximum forbidden band width of E.
g3, a-3i: Indicates that the forbidden band width of the H layer 2304 is Egl.

The multiplication factor of this device was about 10 times or more when a bias of 10V was applied.

Further, even when the wavelength was changed for light having a wavelength of 700 nm or less, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was as low as about 0.1 nA/cm'' or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 2302, and was fast.

(Example 18) This example, like Example 3, is shown in Example 16.

The present inventors have already developed a photoelectric conversion device in Japanese Patent Application Laid-Open No. 63-278.
This is an embodiment in which the present invention is stacked on the scanning circuit and readout circuit proposed in Japanese Patent No. 269.

FIG. 22 is a schematic cross-sectional view of the vicinity of the light receiving section of this embodiment.

The configuration of the photoelectric conversion device of this example is shown in FIG. 5(a).
In the photoelectric conversion device, C is present between the light absorption layer and the multiplication layer.
The structure is the same except that an r electrode (Cr electrode 2414 in FIG. 22) is provided, and its operation is also the same as that of the photoelectric conversion device in FIG. 5(a), so a description thereof will be omitted. .

Further, in the embodiment described above, an example of a circuit according to the invention of the present inventors has been shown, but as in the third embodiment, a -numerical photoelectric conversion device may also be used.

(Example 19) In this example, a light absorption layer having a forbidden band width Egl and absorbing light, an n conductivity type layer having a forbidden band width Eg5, and multiplying carriers generated by absorbing light. The minimum forbidden band width Eg2,
A multiplication layer formed by laminating one or more layers having a step-back structure in which the maximum forbidden band width Eg3 continuously changes, and a layer formed by sequentially laminating layers are sandwiched between charge injection blocking layers. This is how it was done.

In this example, a layer formed by sequentially laminating a light absorption layer, an n-conductivity type layer, and a multiplication layer is sandwiched between charge injection blocking layers, so that the light absorption layer and the multiplication layer are sandwiched between the charge injection blocking layer. placed in n
Since the conductivity type layer functions as a reverse bias layer,
Carriers generated in the light absorption layer are transported smoothly, and high-speed response similar to that of a photodiode without a multiplication layer can be obtained. Further, by reducing the forbidden band width of the n-conductivity type layer, the incidence of light into the multiplication layer is reduced, so that fluctuations in the multiplication factor can be suppressed.

In this example, high sensitivity to visible light can be achieved by setting the forbidden band width Egl of the light absorption layer to a forbidden band width specifically corresponding to visible light.

Furthermore, in this embodiment, by selecting the number of layers of the step-back structure, an amplification factor of 2 or more can be obtained and low noise can be achieved.

In addition, the light absorption layer of the photoelectric conversion device of this example, the n-conductivity type layer,
The components such as the multiplication layer and the charge injection blocking layer are made of at least Si.
If it is made of a single crystal containing atoms, it becomes possible to easily control the forbidden band width and perform low-temperature stacking, thereby solving various problems caused by stacking.

Hereinafter, a nineteenth embodiment of the present invention will be described using FIG. 23 and FIGS. 24(a) and (b).

This example is the same as Example 1 except that a layer formed by sequentially laminating a light absorption layer, an n-conductivity type layer, and a multiplication layer is sandwiched between charge injection blocking layers. Since the explanation is redundant, some of the explanation is omitted.

FIG. 23 is a schematic vertical cross-sectional structural diagram showing a nineteenth embodiment of the photoelectric conversion device of the present invention.

In FIG. 23, 2501 is a Cr electrode, 2502 is an n-type a-5i with a thickness of approximately 500 mm to prevent hole injection.
+-xGex:H charge injection blocking layer, 2503 is a-3it□Gex:H~a for carrier multiplication
-3L+-yCy: multiplication region with changed H composition,
2504 is an n layer for preventing light from entering the multiplication region and strengthening the internal electric field of the light absorption layer to improve the movement of carriers.
Type a-3i+-xGex: 8 layers, 2505 is a-Si with a thickness of about 2 μm to absorb light and generate carriers:
A light absorption layer 2506 is made of H, a charge injection blocking layer made of p-type a-3i:H having a thickness of approximately 100 nm for blocking electron injection, and 2507 is a transparent electrode mainly made of indium oxide.

A Cr electrode 2501 and a transparent electrode 2507 are formed by EB evaporation, and a charge injection blocking layer 2502. Multiplication region 2503, n
Type a-3L+-xGex: 8 layers 2504. Light absorption! 25
The amorphous layers of 05 and charge injection blocking layer 2506 were formed by plasma CVD. The raw material gas for forming the amorphous layer is the charge injection blocking layer 2502 and the n-type a-3it-++.
Gem: 8 layers 2504 are SiH4, GeH4°PH3,
H, multiplication region 2503 is SiH4, GeH4, CH.

H2 was used for the light absorption layer 2505, 5IH4, H2 was used for the light absorption layer 2505, and SiH4, BaH, Hg was used for the charge injection blocking layer 2506.

The multiplication region 2503 consists of three layers 2511, 2512, and 2513 with a thickness of 200 layers in which the gas flow rates of CH4 and GeH4 among the source gases are continuously changed. The gas flow rate of the n-type a-Si 1-xGex:8 layer 2504 is slightly lower than that of the charge injection blocking layer 25o2.

The energy band structure of the photoelectric conversion device of the 19th embodiment shown in FIG. 23 can be theoretically shown in FIGS.
).

FIG. 24(a) shows the energy output when the photoelectric conversion device of the 19th embodiment is in a non-biased state, and FIG. 24(b) shows the energy output when the photoelectric conversion device of the 19th embodiment is in a biased state for carrier multiplication. It is an energy empire.

Figure 24 (a) and (b) are n-type a-Sx+-xGex
:The forbidden band width of the 8th layer 2601 is Hg4, a-3i+-xG
ex:H~a-3i+-yCy:H composition change layer 2611
, 2612, 2613 multiplication region 26 consisting of three layers.
The minimum forbidden band width of o2 is Hg2, the maximum forbidden band width of multiplication region 26o2 is Hg3, the forbidden band width of n-type a-Sx+-++Gex:H layer 26o3 is Hg5, a-Si: 8 layers 260
The forbidden band width of 4 is Egl, and the forbidden band width of p-type a-3i:8 layer 26o5 is Ego.

In addition, in Fig. 24(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen from Fig. 24(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Composition change layer 2611.26122613 created here
The layer giving the maximum forbidden band width Eg3 is a-3i+-, Cy:H, with a C composition ratio y of about 0.4, and Eg3
was about 2.3 eV.

Also, a-3L+, ++Gex: H layer 2601 and n-type a-3it-xGex: Ge1Jl of 8 layer 2603
The ratio X is approximately 0.6, and the forbidden band width Eg4 is approximately 1.3e.
It was V. Composition change layer 2611, 2612.2613
The layer giving the minimum forbidden band width Eg2 is also a−SL+−
*Gei+: H layer, and Eg2 was also about 1.3 eV. 2604.2605 a-3i: The forbidden band widths Egl and Ego of the H layer were both about 1.8 eV.

Furthermore, the light absorption coefficient of the light absorption layer 2603 is at a wavelength of 400 nm.
About I x 105 layers m-' or more for light of m, wavelength 7
It is about 5×10 ″cm−′ or more for light of 0.00 nm, indicating that visible light can be sufficiently absorbed.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Further, for light with a wavelength of 700 nm or less, there was no change in the multiplication factor even if the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 1 nA/am2 or less when a bias of 10 μm was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 2602, and was fast.

Further, in this embodiment, the thickness of the light absorption layer is approximately 2 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 1.8 eV here, but it is also possible to control Eg2 by changing the H2 gas flow rate so as to obtain desired spectral sensitivity characteristics.

(Embodiment 20) Hereinafter, a 20th embodiment of the present invention will be described using FIG. 25.

FIGS. 25(a) and 25(b) are ideally assumed energy band structure diagrams of the 20th embodiment of the present invention.

FIG. 25(a) shows the energy band when the photoelectric conversion device of the 20th embodiment is in an unbiased state, and FIG. 25(b) shows the energy band when the photoelectric conversion device of the 20th embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

In FIG. 25(a), 2701 is n of the forbidden band width Eg4'
Type a-3it-yCy: p-type a-3L + -yCy with H layer and 2705 having forbidden band width Ego':
It is the same as FIG. 24(a) except that it is an H layer and that 2703 is an n-type aSi:H layer with a forbidden band width Eg5', and a-3it-xGex:H~a-Sit-
The minimum forbidden band width of the multiplication region 2702 consisting of three layers of yCy:H composition change layers 2711, 2712, and 2713 is Eg2
, the maximum forbidden band width is Eg3, and the forbidden band width of the a-Si:8 layer 2704 is Egl.

The multiplication factor of this device was approximately 10 times or more when a bias of 10 μ was applied.

Further, even when the wavelength was changed for light having a wavelength of 700 nm or less, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was as low as about 0.1 nA/cm 2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 2702, and was fast.

(Example 21) Similar to Example 3, in this example, the photoelectric conversion device shown in the 19th example was already developed by the present inventors in Japanese Patent Application Laid-Open No. 63-278.
This is a layered structure on the scanning circuit and readout circuit proposed in Japanese Patent No. 269.

Note that the configuration of the photoelectric conversion device is that in the photoelectric conversion device shown in FIG. 5(a), an n-conductivity type layer (
The Cr electrode 2414 in FIG. 22 is an n-conductivity type layer a-3.
The structure is the same except that L+-xGex (same as the one replaced with H) is provided, and its operation is the same as that of the photoelectric conversion device of FIG. 5(a), so a description thereof will be omitted.

Note that in the embodiment described above, an example of a circuit according to the invention of the present inventors was shown, but as in the third embodiment, a general photoelectric conversion device may be used.

(Example 22) This example consists of a light absorption layer that has a forbidden band width Egl and absorbs light, an n conductivity type layer that has a forbidden band width Eg5, and multiplies carriers generated by absorbing light. The minimum forbidden band width Eg2,
A multiplication layer formed by laminating one or more layers having a step-back structure in which the maximum forbidden band width Eg3 continuously changes, and a layer formed by sequentially laminating layers are sandwiched between charge injection blocking layers. The forbidden band width Egl of the light absorption layer and the forbidden band width Eg5 of the n41[type layer]
are approximately equal.

In this example, a layer formed by sequentially laminating a light absorption layer, an n-conductivity type layer, and a multiplication layer is sandwiched between charge injection blocking layers, so that the light absorption layer and the multiplication layer are sandwiched between the charge injection blocking layer. The n-conductivity type layer is given the function of a reverse bias layer to smoothly transport carriers generated in the light-absorbing layer, and furthermore, the forbidden band width Egl of the light-absorbing layer and the forbidden band of the n-conductivity type layer are By making the widths Eg5 substantially equal, band mismatching between the light absorption layer and the multiplication layer and various problems caused by it are resolved, and carrier movement in the light absorption layer due to the generation of interface states, etc. This method prevents a decrease in high-speed response caused by problems with the photoresponse, and provides high-speed response similar to that of a photodiode without a multiplication layer, while at the same time adjusting the forbidden band width Egl of the light absorption layer to correspond specifically to infrared light. By setting the forbidden band width, high sensitivity to infrared light can be achieved. Furthermore, in this embodiment, the amount of light incident on the multiplication layer is reduced, and the variation in the multiplication factor due to the light incident on the multiplication layer is reduced.

Further, in this embodiment, by selecting the number of calendars with a step-back structure, an amplification factor of 2 or more can be obtained and low noise can be achieved.

In addition, the light absorption layer of the photoelectric conversion device of this example, the n-conductivity type layer,
The components such as the multiplication layer and the charge injection blocking layer are made of at least Si.
If it is made of a single crystal containing atoms, it becomes possible to easily control the forbidden band width and perform low-temperature stacking, thereby solving various problems caused by stacking.

Hereinafter, using FIG. 26 and FIGS. 27(a) and (b), the 22nd light absorption layer 17'l +y of the present invention will be described. A layer formed by sequentially stacking layers is sandwiched between charge injection blocking layers, and the forbidden band width Egl of the light absorption layer and the forbidden band width E of the n conductivity type layer are
This embodiment is the same as in Example 1 except that g5 is approximately equal to that of Example 1, and since the explanation will be repeated, a part of the explanation will be omitted.

FIG. 26 is a schematic vertical cross-sectional structural diagram showing a twenty-second embodiment of the photoelectric conversion device of the present invention.

In FIG. 26, 2801 is a Cr electrode, 2802 is an n-type layer-3i with a thickness of approximately 500 nm for blocking hole injection.
A charge injection blocking layer 2803 is made of t-xGex:H, and 2803 is a-3i+-++Gex:H for carrier multiplication.
~a-3i+□Cy: Multiplication region with changed composition of H,
2804 is a layer that prevents light from entering the multiplication region and strengthens the internal electric field of the light absorption layer to improve the movement of carriers.
-3it-xGex: H layer, 2805 is a-Si + with a thickness of about 1 μm to absorb light and generate carriers.
-xGex: A light absorption layer made of H, 2806 is a p-type layer with a thickness of about 100 mm to prevent electron injection -3i +
-, Gex: Charge injection blocking layer made of H, 2807
is a transparent electrode mainly made of indium oxide.

A Cr electrode 2801 and a transparent electrode 2807 are formed by EB evaporation, and a charge injection blocking layer 2802. Multiplication area 2803,
a-5i+-++Gex:H layer 2804. Light absorption layer 2
The amorphous layers 805 and charge injection blocking layer 2806 were formed by plasma CVD. The raw material gas for forming the amorphous layer includes the charge injection blocking layer 2802 and a-3i+-xGe,
l: 1 layer 2804 is si)+4. GeH4, PHi,
H2, multiplication region 2803 is SiH4, GeH4, CH
, H2, the light absorption layer 2805 is SiH4, GeH4, Hz
, the charge injection blocking layer 2806 is made of SiH4, GeH4, Bz
Hs and Hz were used.

The multiplication region 2803 contains CH and Ge1 (4
It consists of three layers, 2811, 2812, and 2813, with a thickness of 200 layers and a composition change layer in which the gas flow rate is continuously changed. a-3i+-++Gex: PH of H layer 2804
, is made slightly smaller than that of the charge injection blocking layer 2802.

The energy band structure of the photoelectric conversion device of the 22nd embodiment shown in FIG. 26 is ideally assumed to be as shown in FIGS. 27(a) and (bl).

FIG. 27(a) shows the energy band when the photoelectric conversion device of the 22nd embodiment is in an unbiased state, and FIG. 27(b) shows the energy band when the photoelectric conversion device of the 22nd embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

Figure 27 (a) and (b) are n-type a-3it-xGex
:The forbidden band width of 8 layers 2901 is Eg4, a-3i+ -x
Gex:H~a-3il-yCy:H The minimum forbidden band width of the multiplication region 2902 consisting of three layers of composition change layers 2911, 2912, and 2913 is Eg2, and the multiplication region 29
The maximum forbidden band width of 02 is Eg3, n type a-3i 1-xG
ex: Forbidden band width of 1st layer 2903 is Eg5, a-Si+
-*Gex: H2904 forbidden band width is Egl, p type a-
3i+-*Gex: Indicates that the forbidden band width of the H layer 2905 is Ego.

In addition, in Fig. 27(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen from Fig. 27(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Composition change layer 2911.29122913 created here
Among them, the layer giving the maximum forbidden band width Eg3 is a-3i+ -yCy:H with a C composition ratio y of about 0.4, and Eg
3 was about 2.3 eV.

Also, a-3t+-++Gex: 8 layers 2901.2
The Ge composition ratio X of 903°2904.2905 is approximately 0.6
and the forbidden band width Eg4. Both Egl and Ego are about 1.
It was 3 eV. Composition change layer 2911, 291229■
The layer giving the minimum forbidden band width Eg2 of 3 is also a-St+
-xGex: 8 layers, and Eg2 was also about 1.3 eV.

Furthermore, the light absorption coefficient of the light absorption layer 2903 is at a wavelength of 800 nm.
About I x 105crtV' or more for light with a wavelength of m, about 2 x 10 'cm-' for light with a wavelength of 11000n
The above results indicate that infrared light can be absorbed sufficiently.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Further, for light having a wavelength of 11,000 nm or less, there was no change in the multiplication factor even if the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 10 nA/cm'' or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 2902, and was fast.

Further, in this embodiment, the thickness of the light absorption layer is approximately 1 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 1.3 eV here, but it is also possible to control Eg2 by changing the Ge composition ratio so as to obtain desired spectral sensitivity characteristics.

(Embodiment 23) Hereinafter, a 23rd embodiment of the present invention will be described using FIG. 28.

FIGS. 28(a) and 28(b) are ideally assumed energy band structure diagrams of the 23rd embodiment of the present invention.

FIG. 28(a) shows the energy band when the photoelectric conversion device of the 23rd embodiment is in an unbiased state, and FIG. 28(b) shows the energy band when the photoelectric conversion device of the 23rd embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

In FIG. 28(a), 3001 is n of the forbidden band width Eg4
Type a-3i1-yCy: Same as FIG. 27(a) except that it has 8 layers, and a-3i+-++Gex:
H~a-3i I-yCy: H composition change layer 3011
, 3012, 3013 multiplication region 3002 consisting of three layers.
The minimum forbidden band width is Eg2, the maximum forbidden band width is Eg3, the forbidden band width of n-type a-3t +-xGem: 8 layers 3003 is Eg5, a-3i+-xGex: the forbidden band width of H layer 3004 is Egl, p-type a-3z+-++Gex:H3005
This shows that the forbidden band width of is EgO.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Further, even when the wavelength was changed for infrared light having a wavelength of 11,000 nm or less, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was reduced to about 10 nA/cm2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin-type photoelectric conversion device without the multiplication layer 3002, and was fast.

(Example 24) Similar to Example 3, in this example, the photoelectric conversion device shown in Example 22 was already developed by the present inventors in Japanese Unexamined Patent Publication No. 63-2782.
This structure is stacked on the scanning circuit and readout circuit proposed in Japanese Patent No. 69.

The structure of the photoelectric conversion device is that in the photoelectric conversion device of FIG. 5(a), the a-Si:8 layer 714 serving as the light absorption layer is replaced with the a-5x + -xGex:8 layer serving as the light absorption layer. , p-type a-3i:l (layer 715 is a light-absorbing layer of p-type a-3i+-xGex:8 layer), and an n-conductivity type layer (see FIG. Cr electrode 24 inside
a-5i + −++Gex where 14 is an n-conductivity type layer:
The photoelectric conversion device is the same as the one shown in FIG.

In addition, in the embodiment described above, an example of the circuit according to the invention of the present inventors was shown, but it is the same as in the third embodiment that the 6-inch photoelectric conversion device may be used.

(Example 25) This example includes a light absorption layer having a forbidden band width Egl and absorbing light, an n conductivity type layer having a forbidden band width Eg5, and multiplying carriers generated by absorbing light. The minimum forbidden band width Eg2,
A multiplication layer formed by laminating one or more layers having a step-back structure in which the maximum forbidden band width Eg3 continuously changes, and a layer formed by sequentially laminating layers are sandwiched between charge injection blocking layers. In addition, the forbidden band width Eg of the light absorption layer
1 exceeds the forbidden band width Eg5 of the n-conductivity type layer.

In this example, a layer formed by sequentially laminating a light absorption layer, an n-conductivity type layer, and a multiplication layer is sandwiched between charge injection blocking layers, so that the light absorption layer and the multiplication layer are sandwiched between the charge injection blocking layer. The n-conductivity type layer is given the function of a reverse bias layer to smoothly transport carriers generated in the light-absorbing layer, and further, the forbidden band width Egl of the light-absorbing layer is equal to the forbidden band of the n-conductivity type layer. By making the width exceed Eg5, band mismatching between the light absorption layer and the multiplication layer and various problems caused by it are resolved, and obstacles to carrier mobility in the light absorption layer due to the generation of interface states etc. It is possible to prevent a decrease in high-speed response caused by factors such as
High sensitivity to ultraviolet light can be achieved by setting l to a forbidden band width that specifically corresponds to ultraviolet light.

Further, in this embodiment, when the forbidden band width Eg5 of the n-conductivity type layer is a narrow forbidden band width, the incidence of light into the multiplication layer is reduced, so that fluctuations in the multiplication factor can be reduced.

Furthermore, in this embodiment, by selecting the number of calendars in the step-back structure, an amplification factor of 2 or more can be obtained, and low noise can be achieved.

Further, the components such as the light absorption layer, the n-conductivity type layer, the multiplication layer, and the charge injection blocking layer of the photoelectric conversion device of this example are at least S
If it is made of a single crystal containing i atoms, it becomes possible to easily control the forbidden band width and perform low-temperature stacking, thereby solving various problems caused by stacking.

Hereinafter, a twenty-fifth embodiment of the present invention will be described using FIG. 29 and FIGS. 30(a) and (b).

In this example, a layer formed by laminating a light absorption layer, an n-conductivity type layer, and a multiplication layer in this order is sandwiched between charge injection blocking layers, and the forbidden band width of the light absorption layer is This example is the same as Example 1 except that Egl exceeds the forbidden band width Eg5 of the n-conductivity type layer, and since the description will be redundant, some of the description will be omitted.

FIG. 29 is a schematic vertical cross-sectional structural diagram showing a twenty-fifth embodiment of the photoelectric conversion device of the present invention.

In Fig. 29, 3101 is a Cr electrode, 3102 is an n-type a-5x with a thickness of approximately 500 mm to prevent hole injection.
+-xGeX:H charge injection blocking layer, 3103 is a-Si+□Gex:H~a for carrier multiplication
-3i +-U C,: Multiplication region where the composition of H is changed,
3104 is an n layer for preventing light from entering the multiplication region and strengthening the internal electric field of the light absorption layer to improve the movement of carriers.
Type a-3i+-++Gex: H layer, 3105 is a-3L with a thickness of about 1 μm to absorb light and generate carriers.
, -yC, :H, 3106 is a p-type a-3L+-y with a thickness of about 100 to prevent electron injection.
A charge injection blocking layer 3107 made of Cy:)I is a transparent electrode mainly made of indium oxide.

A Cr electrode 3101 and a transparent electrode 3107 are formed by EB vaporization, and a charge injection blocking layer 3102. Multiplication region 3103, n
Type a-3it-xGe++: 8 layers 3104. Light absorption layer 3
The amorphous layers 105 and charge injection blocking layer 3106 were formed by plasma CVD. The source gas for forming the amorphous layer is the charge injection blocking layer 3102 and the n-type a-3i + −
xGex: 8 layers 3104 are SiH, GeH4, PH3
°H2, multiplication region 3103 is SiH4, GeH4, CH
4, Hz, the light absorption layer 3105 is SiH4, CH4, H
2. Charge injection blocking layer 3106 is SiH4, CH4, B2
Hs and Hz were used.

The multiplication region 3103 consists of three layers of composition change layers 3111, 3112, and 3113 with a thickness of 200 layers in which the gas flow rates of CH4 and GeH4 among the source gases are continuously changed.
There is. N-type a-3i+ -xGex: 8 layers 310
The gas flow rate at pH 4 is slightly smaller than that of the charge injection blocking layer 3102.

The energy band structure of the photoelectric conversion device of the 25th embodiment shown in FIG. 29 is theoretically as shown in FIGS.
).

FIG. 30(a) shows the energy band when the photoelectric conversion device of the 25th embodiment is in an unbiased state, and FIG. 30(b) shows the energy band when the photoelectric conversion device of the 25th embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

Figure 30 (a) and (b) are n-type a-3i + -xG
ex: Forbidden band width of 8 layers 3201 is Eg4, a-3L+
-xGex:H~a-3Lt-yC,:H composition change layer 3
The minimum forbidden band width of the multiplication region 3202 consisting of three layers of 211, 3212, and 3213 is Eg2, and the multiplication region 3202
The maximum forbidden band width of Eg3, n-type a-3i + -xGe
x: Forbidden band width of 8 layers 3203 is Eg5, a-3i 1
-, C,: The forbidden band width of the 8-layer 3204 is Egl, p-type a-
3it-yCy: Indicates that the forbidden band width of the 8-layer 3205 is EgO.

In addition, in Figure 30(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen from Figure 30(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

A-5in-ycy created here: 8 layers 3204.3
The C composition ratio y of 203 is about 0.4, and the forbidden band width Eg1
, EgO were both about 2.3 eV. Composition change layer 3
Maximum forbidden band width Eg among 211, 3212.3213
The layer giving 3 is also a-3it-yCy: H, and E
g3 was approximately 2.3 eV.

Also, a-3it-xGex: 8 layers 3201 and 3
The Ge composition ratio X of 203 is approximately 0.6, and the forbidden band width Eg
4 was approximately 1.3 eV. Composition change layers 3211, 32
The layer giving the minimum forbidden band width Eg2 of 12.3213 was also the a-Si+ -xGex:H layer, and Eg2 was also about 1.3 eV.

Furthermore, the light absorption coefficient of the light absorption layer 3203 is at a wavelength of 540 nm.
Approximately 4 x 10"cm-' for light of m, wavelength 35
It is about 3 x 10 'Cm-' or more for light of 0 nm, and ultraviolet light can be sufficiently absorbed.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Further, for ultraviolet light having a wavelength of 400 nm or less, there was no change in the multiplication factor even if the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin-type photoelectric conversion device without the multiplication layer 3202, and was fast.

In this embodiment, the thickness of the light absorption layer is approximately 1 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 2.3 eV here, but it is also possible to control Eg2 by changing the C composition ratio so as to obtain desired spectral sensitivity characteristics.

(Embodiment 26) Hereinafter, a 26th embodiment of the present invention will be described using FIG. 31.

FIGS. 31(a) and 31(b) are ideally assumed energy band structure diagrams of the twenty-sixth embodiment of the present invention.

FIG. 31(a) shows the energy output when the photoelectric conversion device of the 26th embodiment is in a non-biased state, and FIG. 31(b) shows the energy output when the photoelectric conversion device of the 26th embodiment is in a biased state for carrier multiplication. It is an energy empire.

In FIG. 31(a), 3301 is n of the forbidden band width Eg4'
Type a-3i+-, Cy: 8 layers, and 3303 is n-type a-3L + -yCy with forbidden band width Eg5':
It is the same as FIG. 30(a) except that there are 8 layers,
The minimum forbidden band width of the multiplication region 3302 consisting of three layers of a-3i+-xGex:H to a-Sit-yCy:H composition change layers 3311.3312.3313 is Eg2, the maximum forbidden band width is Eg3, and a-Sit. -yCy: Forbidden band width of 8 layers 3304 is Egl, p-type a-3i+-, C,: 8 layers 3305
This shows that the forbidden band width of is EgO.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied. Also, there was no change in the multiplication factor even when the wavelength was changed for ultraviolet light with a wavelength of 400 nm or less.

Furthermore, the leakage current in the dark was as low as about 0.1 nA/cm 2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 3302, and was fast.

(Example 27) Similar to Example 3, in this example, the photoelectric conversion device shown in the 25th example was already developed by the present inventors in Japanese Patent Application Laid-Open No. 63-278.
This is a layered structure on the scanning circuit and readout circuit proposed in Japanese Patent No. 269.

The structure of the photoelectric conversion device is that in the photoelectric conversion device of FIG. 5(a), the a-3i:8 layer 714 serving as the light absorption layer is replaced with the a-3i +□C,:H layer serving as the light absorption layer. and p
Type a-3i: P type a-3i in which the H layer 715 is a light absorption layer
+-, C,: 8 layers, and an n4 conductivity type layer (the Cr electrode 2414 in FIG. 22 is an n conductivity type layer) between the light absorption layer and the multiplication layer.+□Gex: The photoelectric conversion device is the same as the one shown in FIG.

In the embodiment described above, an example of the circuit according to the invention of the present inventors was shown, but as in the third embodiment, a -numerical photoelectric conversion device may also be used.

(Example 28) In this example, a light absorption layer having a forbidden band width Egl and absorbing light, an n conductivity type layer having a forbidden band width Eg5, and multiplying carriers generated by absorbing light. The minimum forbidden band width Eg2,
A multiplication layer formed by laminating one or more layers having a step-back structure in which the maximum forbidden band width Eg3 continuously changes, and a layer formed by sequentially laminating layers are sandwiched between charge injection blocking layers. In addition, the forbidden band width Eg of the light absorption layer
1 changes so that it becomes larger continuously from the side of one of the charge injection blocking layers laminated on the light absorption layer, and the forbidden band width Eg1 of the light absorption layer adjacent to the n conductivity type layer is an n conductivity type layer. The forbidden band width Eg5 or more is set.

In this example, a layer formed by sequentially laminating a light absorption layer, an n-conductivity type layer, and a multiplication layer is sandwiched between charge injection blocking layers, so that the light absorption layer and the multiplication layer are sandwiched between the charge injection blocking layer. The n-conductivity type layer formed on the light absorbing layer functions as a reverse bias layer to smoothly transport carriers generated in the light absorbing layer, and furthermore, the forbidden band width Egl of the light absorbing layer is adjusted to It changes so that it becomes larger continuously from the charge injection blocking layer side,
By setting the forbidden band width Egl of the light absorption layer adjacent to the n-conductivity type layer to be equal to or larger than the forbidden band width Eg5 of the n-conductivity type layer, band mismatch between the light absorption layer and the multiplication layer and the resulting band mismatch can be avoided. Various problems have been solved, and it is possible to prevent a decrease in high-speed response caused by obstacles to carrier mobility in the light absorption layer due to the generation of interface states, etc. At the same time as high-speed responsiveness is obtained, high sensitivity from visible light to ultraviolet light can be achieved by making the forbidden band width Egl of the light absorption layer particularly compatible with visible light to ultraviolet light. I can do it.

Furthermore, in this embodiment, when the forbidden band width Eg5 of the n-conductivity type layer is narrow, the incidence of light into the multiplication layer is reduced, so that the variation in the multiplication factor due to the light incidence on the multiplication layer is reduced. can be reduced.

Further, in this embodiment, by selecting the number of layers of the step-back structure, an amplification factor of 2 or more can be obtained and low noise can be achieved.

Further, the components such as the light absorption layer, the n-conductivity type layer, the multiplication layer, and the charge injection blocking layer of the photoelectric conversion device of this example are at least S
If it is made of a single crystal containing i atoms, it becomes possible to easily control the forbidden band width and perform low-temperature stacking, thereby solving various problems caused by stacking.

Hereinafter, a twenty-eighth embodiment of the present invention will be described using FIG. 32 and FIGS. 33(a) and (b).

In this example, a layer formed by sequentially laminating a light absorption layer, an n-conductivity type layer, and a multiplication layer is sandwiched between charge injection blocking layers, and the forbidden band width Eg1 of the light absorption layer is as follows. The forbidden band width Eg1 of the light absorption layer adjacent to the n-conductivity type layer changes so that it becomes larger continuously from the side of one of the charge injection blocking layers laminated on the light-absorption layer, and the forbidden band of the n-conductivity type layer It is the same as in Example 1 except that the width is Eg5 or more, and since the description will be repeated, some of the description will be omitted.

FIG. 32 is a schematic vertical cross-sectional structural diagram showing a twenty-eighth embodiment of the photoelectric conversion device of the present invention.

In Fig. 32, 3401 is a Cr electrode, 3402 is an n-type a-3x with a thickness of approximately 500 mm to prevent hole injection.
+-++Gex: Charge injection blocking layer made of H, 3403
is a-3i + -xGex for carrier multiplication
:H~a-Sit□C: A multiplication region in which the composition of H is changed, 3404 is in order to prevent light from entering the multiplication region and strengthen the internal electric field of the light absorption layer to improve the movement of carriers. n-type a-Sz+-xGex: 8 layers, 3405 is an a-Sz+-xGex with a thickness of about 2 μm to absorb light and generate carriers.
3i:H-a-3it-yC, :H light absorption layer with a changed composition; 3406 is a charge injection blocking layer made of p-type a-3i:H with a thickness of approximately 100 nm for blocking electron injection; , 3407 are transparent electrodes mainly made of indium oxide.

A Cr electrode 3401 and a transparent electrode 3407 are created by EB evaporation, and a charge injection blocking layer 3402. Multiplication region 3403, n
Type a-3i+-++Gex: H layer 3404. Light absorption layer 3
The amorphous layers 405 and charge injection blocking layer 3406 were formed by plasma CVD. The raw material gas for forming the amorphous layer is the charge injection blocking layer 3402 and the n-type a-3i + −
xGex: 8 layers 3404 are SiH4, GeH4, PH
a.

H2, multiplication region 3403 is SiH+, GeH4, CH
+, H2, light absorption layer 3405 is SiH4, CH4, H2
, the charge injection blocking layer 3406 is SiH4, BJs, H2
was used.

The multiplication region 3403 consists of three layers 3411, 3412, and 3413 with a thickness of 200 layers in which the gas flow rates of CH4 and GeH4 among the source gases are continuously changed. The gas flow rate of the n-type a-3i+ -xGex:8 layer 3404 is slightly lower than that of the charge injection blocking layer 3402. The light absorption layer 3405 is formed by continuously changing the flow rate of CH4 among the source gases.

The energy band structure of the photoelectric conversion device of the 28th embodiment shown in FIG. 32 is ideally as shown in FIGS. 33(a) and (b).
) is condemned.

FIG. 33(a) shows the energy consumption when the photoelectric conversion device of the 28th embodiment is in a non-biased state, and FIG. 33(b) shows the energy consumption when the photoelectric conversion device of the 28th embodiment is in a biased state for carrier multiplication. It is an energy empire.

Figures 33(a) and 33(b) show n-type a-3it-, Ge,
:The forbidden band width of 8 layers 3501 is Eg4, a-3i+-xG
em:H-a-3i+-yCy:H composition change layer 3511
, 3512, 3513 multiplication region 35 consisting of three layers.
The minimum forbidden band width of o2 is Eg2, the maximum forbidden band width of the multiplication region 3502 is Eg3, the forbidden band width of the n-type a-3i +-xGeX:H layer 3503 is Eg5, a-3i:H~a-3i
+-yCy:H The minimum forbidden band width of the composition change layer 3504 is E
gl, p-type a-3i: The forbidden band width of the H layer 3505 is Ego
It shows that.

Also, in Figure 33(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen in Figure 33(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Composition change layer 3511.35123513 created here
The layer giving the maximum forbidden band width Eg3 was a-3i+-yCy'' with a C composition ratio y of about 0.4, and Eg3 was about 2.3 eV.a-3i:H~a-3L+ -yC
The layer giving the maximum forbidden band width of the y:H composition change layer 3504 was also a-3i+-, C,:H.

Also, a-3x+-xGex: H layers 3501 and 3
The Ge composition ratio X of 503 is approximately 0.6, and the forbidden band width Eg
4 was approximately 1.3 eV. Composition change layers 3511, 35
The layer giving the minimum forbidden band width Eg2 of 12.3513 was also the a-St+-xGex:H layer, and Eg2 was also about 1.3 eV. a-Si+H~a-3it-yC
The layer giving the minimum forbidden band width Egl of the y:H composition change layer 3504 was a-5L:H, and Egl was about 1.8 eV. P-type a-3i: Forbidden band width Ego of H layer 3505
was also about 1.8.

Furthermore, the light absorption coefficient of the light absorption layer 3504 is at a wavelength of 700 nm.
About 6 x 103 cm-' for light of m, wavelength 35
This is about 3 x 10 'cm-' or more for light of 0 nm, and absorption of visible light to ultraviolet light is sufficient.

The multiplication factor of this device was approximately 10 times or more when a bias of 1OV was applied.

Furthermore, there was no change in the multiplication factor even when the wavelength was changed from the visible region to the ultraviolet region with a wavelength of 700 nm or less.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm'' or less when a bias of 10 μm was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 3502, and was fast.

Yes.

Further, in this embodiment, the thickness of the light absorption layer is approximately 2 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 1.8 to 2.3 here.
eV, but Eg2 is controlled by changing the C composition ratio,
It is also possible to obtain desired spectral sensitivity characteristics.

(Embodiment 29) Hereinafter, a 29th embodiment of the present invention will be described using FIG. 34.

FIGS. 34(a) and 34(b) are ideally assumed energy band structure diagrams of the 29th embodiment of the present invention.

FIG. 34(a) shows the energy band when the photoelectric conversion device of the 29th embodiment is in an unbiased state, and FIG. 34(b) shows the energy band when the photoelectric conversion device of the 29th embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

In FIG. 34(a), 3601 is n of the forbidden band width Eg4'
Type a-Si+-yCy:H layer and light absorption layer 3
604 a-3i: The H composition region is wider than that in FIG. 33, and 3603 is an n-type a-3 with a forbidden band width Eg5'.
Figure 33(a) except that it is an i+□C,:H layer.
is the same as a−Sx+−xGex:H”””a−
3l+-yCy:H composition change layer 3611, 3612.3
The minimum forbidden band width of the multiplication region 3602 consisting of three layers of 613 is Eg2, the maximum forbidden band width is Eg3, a-Si:H~a-
3it-ycy: The minimum forbidden band width of the H layer 3604 is Egl
, p-type a-3i: The forbidden band width of the H layer 3605 is Eg.

It shows that.

The multiplication factor of this device was about 10 times or more when a bias of 10V was applied.

Furthermore, even when the wavelength was changed from the visible wavelength region of 700 nm or less to the ultraviolet wavelength region, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was as low as about 1 nA/am'' or less when a bias of 10 V was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 36o2, and was fast.

(Example 30) In this example, similar to Example 3, the photoelectric conversion device shown in Example 28 was already developed by the present inventors in Japanese Patent Application Laid-Open No. 63-27826.
It is laminated on the scanning circuit and readout circuit proposed in Publication No. 9.

Note that the structure of the photoelectric conversion device is that in the photoelectric conversion device of FIG. layer, and between the light absorption layer and the multiplication layer is an n-conductivity type layer (the Cr electrode 2414 in FIG.
+-xGex: Same as the one replaced with H) is provided, and its operation is also shown in Figure 5 (a
) is the same as the photoelectric conversion device, so the explanation will be omitted.

In addition, in the embodiment described above, an example of the circuit according to the invention of the present inventors has been shown, but as in the third embodiment, a photoelectric conversion device for boat fishing may also be used.

(Example 31) In this example, a light absorption layer having a forbidden band width Egl and absorbing light, an n conductivity type layer having a forbidden band width Eg5, and multiplying carriers generated by absorbing light. The minimum forbidden band width Eg2,
A multiplication layer formed by laminating one or more layers having a step-back structure in which the maximum forbidden band width Eg3 continuously changes, and a layer formed by sequentially laminating layers are sandwiched between charge injection blocking layers. At the same time, the forbidden band width Egl of the light absorption layer changes so as to become smaller continuously from the side of one charge injection blocking layer laminated on the light absorption layer, and The forbidden band width Egl of the light absorption layer is n
This is made approximately equal to the forbidden band width Eg5 of the conductivity type layer.

In this example, a layer formed by sequentially laminating a light absorption layer, an n-conductivity type layer, and a multiplication layer is sandwiched between charge injection blocking layers, so that the light absorption layer and the multiplication layer are sandwiched between the charge injection blocking layer. The n-conductivity type layer in the layer is given the function of a reverse bias layer to smoothly transport carriers generated in the light absorption layer, and furthermore, the forbidden band width Egl of the light absorption layer is laminated on the light absorption layer. The forbidden band width Egl of the light absorption layer adjacent to the n conductivity type layer changes so as to become smaller continuously from the charge injection blocking layer side,
By configuring the bandgap Eg5 to be approximately equal to the forbidden band width Eg5 of the n-conductivity type layer, band mismatch between the light absorption layer and the multiplication layer and various problems caused by it are eliminated, and light The absorption layer can prevent a decrease in high-speed response caused by obstacles to carrier running properties, etc.
A high-speed response similar to that of a photodiode without a multiplication layer can be obtained, and at the same time, the forbidden band width Egl of the light absorption layer is set to a band gap corresponding to infrared light to ultraviolet light.
High sensitivity can be achieved from infrared light to ultraviolet light.

Further, in this embodiment, when the forbidden band width Eg5 of the n-conductivity type layer is a narrow forbidden band width, the incidence of light into the multiplication layer is reduced, so that fluctuations in the multiplication factor can be reduced.

Further, in this embodiment, by selecting the number of calendars with a step-back structure, an amplification factor of 2 or more can be obtained and low noise can be achieved.

Further, the components such as the light absorption layer, the n-conductivity type layer, the multiplication layer, and the charge injection blocking layer of the photoelectric conversion device of this example are at least S
If it is made of a single crystal containing i atoms, it becomes possible to easily control the forbidden band width and perform low-temperature stacking, thereby solving various problems caused by stacking.

Hereinafter, a 31st embodiment of the present invention will be described using FIG. 35 and FIGS. 36(a) and (b).

In this example, a layer formed by sequentially laminating a light absorption layer, an n-conductivity type layer, and a multiplication layer is sandwiched between charge injection blocking layers, and the forbidden band of the light absorption layer is The width Egl changes continuously from the side of one charge injection blocking layer laminated on the light absorption layer, and the forbidden band width Egl of the light absorption layer adjacent to the n conductivity type layer is of the n conductivity type. Layer forbidden band width E
This is the same as in Example 1 except that it is configured to be approximately equal to g5, and since the description will be redundant, some of the description will be omitted.

FIG. 35 is a schematic vertical cross-sectional structural diagram showing a 31st embodiment of the photoelectric conversion device of the present invention.

In FIG. 35, 3701 is a Cr electrode, 3702 is an n-type a-3i with a thickness of approximately 500 mm to prevent hole injection.
, -xGex:H, and 3703 is a-sL□Gex:H~a- for carrier multiplication.
3i+□Cy: Multiplication region with changed H composition, 370
4 is an n type a-
3i I-xGex: 1 layer, 3705 is a-5i:H with a thickness of about 1 μm to absorb light and generate carriers.
-a-3i+-, a light absorption layer with a changed composition of Ge:H, 3706 is a charge injection blocking layer made of p-type a-3i:H with a thickness of approximately 100 nm for blocking electron injection, 370
7 is a transparent electrode mainly made of indium oxide.

A Cr electrode 3701 and a transparent electrode 3707 are created by EB evaporation, and a charge injection blocking layer 3702. Multiplication region 3703, n
Type a-3t + -xGex: 8 layers 3704. The amorphous layers of the light absorption layer 3705 and the charge injection blocking layer 3706 were formed by plasma CVD. The source gas for forming the amorphous layer is the charge injection blocking layer 3702 and the n-type a-Si+-
xGex: 8 layers 3704 are SiH4, GeH+, PH
s.

H2, multiplication region 3703 is SiH4, Ge1-14. C
H4, H2, the light absorption layer 3705 is SiH4, GeH4,
Ha, siL, LHs, and L were used for the charge injection blocking layer 3706.

The multiplication region 3703 consists of three layers 3711, 3712, and 3713 with a thickness of 200 layers in which the gas flow rates of CH4 and GeHa among the source gases are continuously changed. n-type a-3i+-xGex: PH of 8 layers 3704
, is made slightly smaller than that of the charge injection blocking layer 3702. The light absorption layer 3705 contains GeH among the raw material gases.
4 by continuously changing the gas flow rate.

The energy band structure of the photoelectric conversion device of the 31st embodiment shown in FIG. 35 is ideally as shown in FIGS. 36(a) and (b).
).

FIG. 36(a) shows the energy consumption when the photoelectric conversion device of the 31st embodiment is in a non-biased state, and FIG. 36(b) shows the energy consumption when a bias is applied to perform carrier multiplication operation. It is an energy empire.

Figures 36(a) and 36(b) show n-type a-St+-xGex:
The forbidden band width of 8 layers 3801 is Eg4, a-Si+-xGe
x: H~a-3x + -yCy: H composition change layer 38
The minimum forbidden band width of the multiplication region 3802 consisting of three layers of 11, 3812, and 3813 is Eg2, the maximum forbidden band width of the multiplication region 3802 is Eg3, n-type a-3i+-xGex: 8
The forbidden band width of the layer 3803 is Eg5, a-3t: H ~ a-3
i1-, Gex: The maximum forbidden band width of the H layer 3804 is Egl
, indicating that the forbidden band width of the p-type a-3i+H layer 38O5 is EgO.

Furthermore, in Fig. 36(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen from Fig. 36(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Composition change Jl created here! 3811.381238
The layer giving the maximum forbidden band width Eg3 out of 13 is a-3it-yCy:H with a C composition ratio y of about 0.4, and E
g3 was approximately 2.3 eV.

Also, a-Si+-xGex: 8 layers 3801 and 3
The Ge composition ratio X of 803 is approximately 0.6, and the forbidden band width Eg
4 was approximately 1.3 eV. Composition change layers 3811, 38
The layer giving the minimum forbidden band width Eg2 of 12.3813 was also an a-Si+-xGex:H layer, and Eg2 was also about 1.3e■. The layer giving the minimum forbidden band width of the 3804 layer was also a-3i+, ++Gex:H. 3
The layer that gives the maximum forbidden band width Egl of the 8o4 layer is a-3i
: H, and Egl was about 1.8 eV. 3
The forbidden band width Ego of the 805 layer was also about 1.8 eV.

Furthermore, the light absorption coefficient of the light absorption layer 3804 is at a wavelength of 400 nm.
About I x 10 'cm-' or more for light with a wavelength of m, and about 2 x 10 'cm-' for light with a wavelength of 11000 nm.
The above results indicate that infrared light, visible light, and ultraviolet light can be sufficiently absorbed.

The multiplication factor of this device was about 10 times or more when a bias of 10V was applied.

Furthermore, for light with a wavelength of 1000 nm or less, there was no change in the multiplication factor even if the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm'' or less when a bias of 10 μm was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 3802, and was fast.

In this embodiment, the thickness of the light absorption layer is approximately 1 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 1.8 to 1.3 here.
eV, but it is also possible to control Eg2 by changing the Ge composition ratio to obtain desired spectral sensitivity characteristics.

(Embodiment 32) Hereinafter, a 32nd embodiment of the present invention will be described using FIG. 37.

FIGS. 37(a) and 37(b) are ideally assumed energy band structure diagrams of the 32nd embodiment of the present invention.

FIG. 37(a) shows the energy band when the photoelectric conversion device of the 32nd embodiment is in an unbiased state, and FIG. 37(b) shows the energy band when the photoelectric conversion device of the 32nd embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

In FIG. 37(a), 3901 is n of the forbidden band width Eg4'.
It is the same as FIG. 36(a) except that it is a type a-5ir-yCy:8 layer and the a-3i:H composition region of the light absorption layer 3904 is wider than that in FIG. 36, a-
The minimum forbidden band width of the multiplication region 3902 consisting of three layers of Si+-xGex:H~a-3it-yCy:H composition change layers 3911, 3912, and 3913 is Eg2, the maximum forbidden band width is Eg3, and n-type a-3i. + -xGex: 8 layers 390
The forbidden band width of 3 is Eg5, a-3i:H~a-3it-x
Ge, :H layer 3904(7) Maximum forbidden layer width 904l,p
Type a-3i: This shows that the forbidden band width of the 8-layer 3905 is EgO.

The multiplication factor of this device was approximately 10 times or more when IOV bias was applied.

Further, even when the wavelength was changed for light having a wavelength of 1000 nm or less, there was no change in the multiplication factor.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm'' or less when a bias of 10 μm was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 3902, and was fast.

(Example 33) Similar to Example 3, in this example, the photoelectric conversion device shown in Example 31 was already developed by the present inventors in Japanese Unexamined Patent Publication No. 63-2782.
This structure is stacked on the scanning circuit and readout circuit proposed in Japanese Patent No. 69.

Note that the configuration of the photoelectric conversion device is as follows: in the photoelectric conversion device of FIG. In addition to the H layer, an n-conductivity type layer (
The Cr electrode 2414 in FIG. 22 is an n-conductivity type layer a-3.
i+-++Gex: Same as the one replaced with H) is provided, and its operation is also the same as that of the photoelectric conversion device of FIG. 5(a), so a description thereof will be omitted.

In addition, in the embodiment described above, an example of the circuit according to the invention of the present inventors has been shown, but as in the third embodiment, a photoelectric conversion device for boat fishing may also be used.

(Embodiment 34) In this embodiment, a plurality of photoelectric conversion sections, a storage means for accumulating electrical signals generated from the photoelectric conversion sections, and an electrical signal generated from the photoelectric conversion sections are scanned. A photoelectric conversion device having a signal output section having at least six means among the reading means for reading out the electrical signals generated from the photoelectric conversion section; a light absorption layer that absorbs light; and a multiplication layer formed by laminating one or more layers of a step-back structure in which the forbidden band width, ie, the minimum forbidden band width Eg2 and the maximum forbidden band width Eg3, multiplies the carriers generated by absorbing light. The photoelectric conversion section is constructed by sandwiching the photoelectric conversion section between a charge injection blocking layer laminated on the light absorption layer and a base body having a charge injection blocking function on which the signal output section is formed.

In this example, a light absorption layer and a multiplication layer are sandwiched between a charge injection blocking layer laminated on the light absorption layer and a substrate having a charge injection blocking function that forms the signal output section, so that photoelectric conversion is possible. Since the substrate having a charge injection blocking function on the multiplication layer side is in a state of being patterned in advance, there is no need to separate the elements by reactive ion etching or the like. Therefore, process problems caused by reactive ion etching, etc. are solved, and the increase in dark current leakage caused by isolation between elements is suppressed, and it has excellent high-speed response, low noise, and high sensitivity, and can be used for light of various wavelengths. It is possible to provide a photoelectric conversion device that can easily be made large in size.

Hereinafter, a 34th embodiment of the present invention will be described using FIG. 38 and FIGS. 39(a) and (b).

Note that, since the characteristic part of the photoelectric conversion device of the present invention is in the photoelectric conversion section, only the photoelectric conversion section will be explained here, and the overall structure and operation of the photoelectric conversion device will be described later.

In the photoelectric conversion section of this embodiment, a light absorption layer and a multiplication layer are sandwiched between a charge injection blocking layer laminated on the light absorption layer and a substrate having a charge injection blocking function that forms the signal output section. Except for this, it is the same as Example 1 and the explanation is duplicated, so
Some explanations are omitted.

FIG. 38 is a schematic vertical cross-sectional structural diagram showing a photoelectric conversion section of the 34th embodiment of the photoelectric conversion device of the present invention.

In FIG. 38, 4001 is an A°C electrode, 4002 is an n-type C-3i substrate that serves as a base with a function of blocking electron injection, and 4003 is an a-3i+- substrate for carrier multiplication.
xGex:H"""a-3ll-yCy: Multiplication region with a changed composition of H, 4004 is a light absorption layer made of a-3i:11 with a thickness of about 2 μm for absorbing light and generating carriers. , 4005 is a charge injection blocking layer made of p-type a-3i:H having a thickness of approximately 100 nm for blocking electron injection, and 4006 is a transparent electrode mainly made of indium oxide.

The Cr electrode 4001 and the transparent electrode 4006 are created by EB evaporation, and the multiplication region 4003. The amorphous layers of the light absorption layer 4004 and the charge injection blocking layer 4005 were formed by plasma CVD. The raw material gas when creating the amorphous layer is in the multiplication region 40.
03 is SiH<, GeH4°CH4,H2, light absorption layer 4004 is 5i)14. )+! , charge injection blocking layer 400
5 is SiH4, BJs,) 1 g was used.

The multiplication region 4003 consists of three layers 4011, 4012, and 4013 with a thickness of 200 layers in which the gas flow rates of CH4 and GeH4 among the source gases are continuously changed.

The energy band structure of the photoelectric conversion device of the 34th embodiment shown in FIG. 38 is ideally as shown in FIGS. 39(a) and (b).
).

FIG. 39(a) shows the energy band when the photoelectric conversion section of the 34th embodiment is in a non-biased state, and FIG. 39(b) shows the energy band when the photoelectric conversion section of the 34th embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same.

Figures 39(a) and 39(b) show an n-type C-3i substrate 41o1
The forbidden band width of CH4, a-5i+-xGex:H~a-
5it-yC, :H composition change layer 4111, 4112.4
The minimum forbidden band width of the multiplication region 4102 consisting of three layers of 113 is CH2, and the maximum forbidden band width of the multiplication region 4102 is CH
3. The forbidden band width of the a-Si:H layer 4103 is Egl, and the forbidden band width of the p-type a-3i:8 layer 4104 is EgO.

Also, in Figure 39(a), there are energy discontinuities at both the conduction band edge and the valence band edge, but when a bias voltage is applied, as can be seen in Figure 39(b), There are almost no barriers due to energy discontinuity in the direction in which the carrier travels, and the carrier's travel performance is not hindered.

Composition change layers 4111, 41124113 created here
The layer giving the maximum forbidden band width Eg3 is a-3i,-,C,:Htet with a C composition ratio y of about 0.4.

, CH3 was about 2.3 eV.

Further, among the composition change layers 4111, 4112, and 4113, the layer providing the minimum forbidden band width Eg2 has a Ge composition ratio X of 0.
6 a-311-++Gex: 8 layers, CH2
is about 1.3 eV, 4103.4104 (7
) The forbidden band widths Egl and Ego of the a-3i biH layer were both about 1.8 eV.

Furthermore, the light absorption coefficient of the light absorption layer 4103 is at a wavelength of 400 nm.
Approximately lx10'c1 or more for light of m, wavelength 700
It is about 5×10 ″am−′ or more for light of nm, and visible region light can be sufficiently absorbed.

The multiplication factor of this device was approximately 10 times or more when a bias of 1 oV was applied.

Furthermore, for light with a wavelength of 700 nm or less, there was no change in the multiplication factor even when the wavelength was changed.

Furthermore, the leakage current in the dark was as low as about 1 nA/cm 2 or less when a bias of 10 V was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 4102, and was fast.

In this embodiment, the thickness of the light absorption layer was approximately 2 μm, but it is sufficient that the thickness is such that the incident light does not pass through the light absorption layer and reach the multiplication layer. This thickness is determined by the light absorption coefficient.

The forbidden band width Eg2 of the light absorption layer is approximately 1.8 eV here, but it is also possible to control Eg2 by changing the H2 gas flow rate so as to obtain desired spectral sensitivity characteristics.

(Embodiment 35) Hereinafter, a 35th embodiment of the present invention will be described using FIG. 40.

FIG. 40 is a schematic vertical cross-sectional structural diagram showing a photoelectric conversion section of the 35th embodiment of the photoelectric conversion device of the present invention.

In FIG. 40, 42o2 is the same as in FIG. 38 except that 42o2 is an n-type polySL layer having a function of blocking electron injection created by low pressure CVD method, and 4201 is an Af
a-5 for f electrode, 4203 to perform carrier multiplication
Multiplication region consisting of three layers 4211, 4212 and 4213 with different compositions of i+-++Ge*:t(-a-3it-yCy:H, thickness for 4204 to absorb light and generate carriers. A light absorbing layer of about 2 μm of a-3i:H, 4205 with a thickness of about 1 to prevent electron injection.
00 p-type a-3i:) a charge injection blocking layer consisting of I;
4206 is a transparent electrode mainly made of indium oxide.

The multiplication factor of this device was about 10 times or more when a bias of 10V was applied.

Furthermore, there was no change in the multiplication factor even when the wavelength was changed for ultraviolet light having a wavelength of 400 nm or less.

Furthermore, the leakage current in the dark was as low as about 0.1 nA/cm 2 or less when IOV bias was applied.

Furthermore, the optical response speed was the same as that of a pin type photoelectric conversion device without the multiplication layer 4202, and was fast.

Next, the overall structure and operation of the photoelectric conversion device of the present invention will be explained.

In the embodiment described below, the photoelectric conversion section shown in embodiment 34 is laminated on the scanning circuit and readout circuit which the present inventors had already proposed in Japanese Patent Laid-Open No. 63-278269.

FIG. 41 is a schematic cross-sectional view of the vicinity of the light receiving section of an embodiment of the photoelectric conversion device of the present invention.

The circuit diagram of one pixel, the equivalent circuit of the entire device, and the block circuit are bipolar transistors 731 and M.
Except for changing the conductivity type of the OS transistor 732.
Since it is equivalent to that shown in FIG. 5(b) and FIG. 5(c), the same reference numerals will be used and the explanation will be given with reference to both figures.

In FIG. 41, a p- layer 430 which becomes a collector region is epitaxially grown on a p-type silicon substrate 4301.
2 is formed, and an n base region 4303 and a p'' emitter region 4304 are formed therein to constitute a bipolar transistor.

The n base region 4303 is separated from adjacent pixels, and a gate electrode 4306 is formed between the horizontally adjacent n base regions with an oxide film 4305 in between. Therefore, an n-channel MOS transistor is configured with adjacent n base regions 4303 as source and drain regions, respectively. Gate electrode 4306 also functions as a capacitor for controlling the potential of n base region 4303.

Furthermore, after forming the insulating layer 4307, the emitter electrode 4
308 is formed.

After that, insulation 94309 is formed, and then a-3i,
, Ge, +H~a-3L, -yC, :H composition change layer 4
321.4322.4323 to form the multiplication region 431
Configure 0. Next, light absorption layer a-3i: 8 layers 4311
A p-type a-3i:8 layer 4312 is formed, and a transparent electrode 4313 for applying a bias voltage to the sensor is formed.
form.

Further, a collector electrode 4314 is ohmically connected to the back surface of the substrate 4301.

Therefore, as shown in FIG. 5(b), the equivalent circuit of one pixel is a bibolar transistor 7 made of crystalline silicon.
An n-channel MOS transistor 73 is connected to the base of 31.
2, a capacitor 733, and a photoelectric conversion device 734 similar to Embodiment 1 are connected, a terminal 735 for applying a potential to the base, a terminal 736 for driving the n-channel MOS transistor 732 and the capacitor 733, and a sensor electrode 737. , an emitter electrode 738, and a collector electrode 739.

Figure 5(c/) is the one pixel self 4 shown in Figure 5(b).
00 is arranged in a 3×3 two-dimensional matrix.

In the figure, a collector electrode 741 of one pixel cell 40
are provided in all pixels, and sensor electrodes 742 are also provided in all pixels. Furthermore, the gate electrodes and capacitor electrodes of the nMO3 transistors are connected to the drive wiring 41743.743'743'' for each row, and are connected to the vertical shift register (V, S, R) 744.

Further, the emitter electrodes are connected to vertical wirings 746, 746', 746'' for signal readout for each column.The vertical wirings 746, 746', 746'' are connected to switches 747, 747' for resetting the charges of the vertical wirings, respectively. 7
47'' and readout switches 750, 750', 750~.Reset switches 747, 747',
The gate electrodes of 747'' are commonly connected to a terminal 748 for applying a vertical line reset pulse, and the source electrodes are commonly connected to a terminal 748 for applying a vertical line reset voltage.
9 are commonly connected. Read switch 750750
'750'' gate electrodes are wires 751 and 75, respectively.
1', 751'' to the horizontal shift register (H, S,
R) 752, and its drain electrode is connected to an output amplifier 757 via a horizontal readout wiring 753. The horizontal readout line 753 is connected to a switch 754 for resetting the charge of the horizontal readout line.

The reset switch 754 is connected to a terminal 755 for applying a horizontal wiring reset pulse and a terminal 756 for applying a horizontal wiring reset voltage.

Finally, the output of amplifier 757 is taken out from terminal 758.

The operation will be briefly explained below using FIG. 41 and FIGS. 5(b) and (c).

The incident light is absorbed by the light absorption layer 4311 in FIG. 41, and the generated carriers are multiplied in the multiplication region 4310.
Accumulated within base region 4303.

When a negative drive pulse output from the vertical shift register in FIG. 5(c) appears on the drive wiring 743, the base potential decreases via the capacitor, and a signal charge corresponding to the amount of light is released from the pixels in the first row. Vertical yarn spring 746, 746', 7
46" respectively.

Next, a scan pulse of 75 is sent from the horizontal shift register 752.
1,751' and 751'', the switches are sequentially turned ON and OFF, and the signal is taken out to the output terminal 758 through the amplifier 757. At this time, the reset switch 754 is niO
It is in the N state, and the residual charge on the horizontal wiring 753 is removed.

Next, vertical line reset switches 747°747', 7
47'' is in the ON state, and the vertical wiring 746, 746',
746" residual charge is removed. Then, when a positive pulse is applied from the vertical shift register 744 to the drive wiring 743, the nMO3 transistor of each pixel in the first row becomes O.
The N state is reached, the base residual charge of each pixel is removed, and the pixel is initialized.

Next, a drive pulse output from the vertical shift register 744 appears on the drive wiring 743', and the signal charges of the pixels in the second row are similarly taken out.

Next, the signal charges of the pixels in the third row are extracted in the same manner.

The device operates by repeating the above operations.

In addition, in the embodiment described above, an example of the circuit according to the invention of the present inventors was shown, but as in the third embodiment, a -a type photoelectric conversion device may also be used.

[Effects of the Invention] As explained above, according to the photoelectric conversion device of the present invention,
A light absorption layer that absorbs light and has a forbidden band width Egl, and a minimum forbidden band width Eg that multiplies carriers generated by absorbing light.
2. A multiplication layer formed by stacking one or more layers of a step-back structure in which the maximum forbidden band width Eg3 continuously changes is sandwiched between charge injection blocking layers. , the following effects can be obtained.

(1) Since the independent light absorption layer is sandwiched between the stepback layer and the charge injection blocking layer, light penetration into the multiplication layer is reduced, and fluctuations in the multiplication factor due to light penetration into the multiplication layer are reduced.

(2) The multiplication layer is a step-back structure layer with a large ΔEc, and electrons are multiplied, and a sufficient multiplication factor can be achieved with low noise. For example, in the example according to the present invention, a multiplication factor of about 10 times could be obtained with three step-back structure layers.

(3) A non-single-crystal material is desirable as a constituent material of an element to which the present invention is applied. Here, the non-single crystal material is a polycrystalline material or an amorphous material, and the amorphous material is
The so-called microcrystalline structure is also included in the scope.

Specifically, amorphous silicon compensated with hydrogen and/or halogen elements (hereinafter referred to as a-3i (H,X))
, amorphous silicon germanium (hereinafter referred to as a-3iGe (
H,X)), amorphous silicon carbide (hereinafter referred to as a)
-3iC (H,
20] [311] It also includes amorphous silicon having crystallinity that has a peak specified by each Miller index.

In this way, since the constituent material of the element is a non-single crystal material, it can be processed at low temperatures (for example, 200 to 30
0°C) and can be easily fabricated on a large-area substrate, and the forbidden band width can be easily controlled and the composition modulated, making it relatively easy to create a multiplication layer with a step-back structure. Heat diffusion etc. are suppressed, a relatively reliable step-back structure can be formed, and problems in laminating multiple layers can be reduced.

(4) Hydrogenated amorphous silicon or the like has a large light absorption coefficient, so the thickness of the light absorption layer can be reduced.

(5) The forbidden band width of the light absorption layer also increases the degree of freedom for the same reason as mentioned in (3) above. A sensitive photoelectric conversion element can be constructed.

As a result of the above-described effects, it is possible to obtain a novel photoelectric conversion device with low noise, high sensitivity, and easy expansion into a large area.

[Brief explanation of drawings]

FIG. 1(a) is a schematic cross-sectional structural diagram showing the structure of the photoelectric conversion device of the present invention, FIG. 1(b) is the energy band surface of the photoelectric conversion device when no bias is applied, and FIG. ) is a diagram showing energy band 9 when the photoelectric conversion device is reverse biased. FIG. 2 is a schematic vertical cross-sectional structural diagram showing a first embodiment of the photoelectric conversion device of the present invention. FIG. 3(a) shows the energy band surface when the photoelectric conversion device of the first embodiment is in an unbiased state, and FIG. 3(b) shows the energy band surface when the photoelectric conversion device of the first embodiment is in a biased state to perform a carrier multiplication operation. is the energy band surface of FIG. 4(a) shows the energy band surface when the photoelectric conversion device of the second embodiment is in an unbiased state, and FIG. 4(b) shows the energy band surface when the photoelectric conversion device of the second embodiment is in a biased state to perform a carrier multiplication operation. is the energy band surface of FIG. 5(a) is a schematic cross-sectional view of the vicinity of the light receiving part of the third embodiment of the photoelectric conversion device of the present invention, FIG. 5(b) is an equivalent circuit diagram of one pixel, and FIG. 5(C) is the main 3 is an equivalent circuit and a block circuit diagram of the entire device. FIG. FIG. 6 is a block diagram showing a configuration when the present invention is applied to a photoelectric conversion device having a general configuration. FIG. 7 is a schematic vertical cross-sectional structural diagram showing a fourth embodiment of the photoelectric conversion device of the present invention. FIG. 8(a) shows the energy band surface when the photoelectric conversion device of the fourth embodiment is in an unbiased state, and FIG. 8(b) shows the energy band surface when the photoelectric conversion device of the fourth embodiment is in a biased state to perform a carrier multiplication operation. is the energy band surface of FIG. 9(a) shows the energy band when the photoelectric conversion device of the fifth embodiment is in an unbiased state, and FIG. 9(b) shows the energy band when the photoelectric conversion device of the fifth embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same. FIG. 10 is a schematic vertical cross-sectional structural diagram showing a seventh embodiment of the photoelectric conversion device of the present invention. FIG. 11(a) shows the energy band when the photoelectric conversion device of the seventh embodiment is in an unbiased state, and FIG. 11(b) shows the energy band when the photoelectric conversion device of the seventh embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same. FIG. 12(a) shows the energy band when the photoelectric conversion device of the eighth embodiment is in an unbiased state, and FIG. 12(b) shows the energy band when the photoelectric conversion device of the eighth embodiment is in a biased state to perform carrier multiplication operation. The energy band is the same. FIG. 13 is a schematic vertical cross-sectional structural diagram showing a tenth embodiment of the photoelectric conversion device of the present invention. FIG. 14(a) shows the energy band when the photoelectric conversion device of the 10th embodiment is in an unbiased state, and FIG. 14(b) shows the energy band when the photoelectric conversion device of the 10th embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same. FIG. 15(a) shows the energy band when the photoelectric conversion device of the eleventh embodiment is in an unbiased state, and FIG. 15(b) shows the energy band when the photoelectric conversion device of the eleventh embodiment is in a biased state to perform carrier multiplication operation. The energy band is the same. FIG. 16 is a schematic vertical cross-sectional structural diagram showing a thirteenth embodiment of the photoelectric conversion device of the present invention. FIG. 17(a) shows the energy band when the photoelectric conversion device of the 13th embodiment is in an unbiased state, and FIG. 17(b) shows the energy band when the photoelectric conversion device of the 13th embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same. FIG. 18(a) shows the energy band when the photoelectric conversion device of the 14th embodiment is in an unbiased state, and FIG. 18(b) shows the energy band when the photoelectric conversion device of the 14th embodiment is in a biased state for carrier multiplication. The energy band is the same. FIG. 19 is a schematic vertical cross-sectional structural diagram showing a 16th embodiment of the photoelectric conversion device of the present invention. FIG. 20(a) shows the energy band when the photoelectric conversion device of the 16th embodiment is in an unbiased state, and FIG. 20(b) shows the energy band when the photoelectric conversion device of the 16th embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same. FIG. 21(a) shows the energy band when the photoelectric conversion device of the 17th embodiment is in an unbiased state, and FIG. 21(b) shows the energy band when the photoelectric conversion device of the 17th embodiment is in a biased state to perform carrier multiplication. The energy band is the same. FIG. 22 is a schematic cross-sectional view of the vicinity of the light receiving section of this embodiment. FIG. 23 is a schematic vertical cross-sectional structural diagram showing a nineteenth embodiment of the photoelectric conversion device of the present invention. FIG. 24(a) shows the energy band when the photoelectric conversion device of the 19th embodiment is in an unbiased state, and FIG. 24(b) shows the energy band when the photoelectric conversion device of the 19th embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same. FIG. 25(a) shows the energy band when the photoelectric conversion device of the 20th embodiment is in an unbiased state, and FIG. 25(b) shows the energy band when the photoelectric conversion device of the 20th embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same. FIG. 26 is a schematic vertical cross-sectional structural diagram showing a twenty-second embodiment of the photoelectric conversion device of the present invention. FIG. 27(a) shows the energy band when the photoelectric conversion device of the 22nd embodiment is in an unbiased state, and FIG. 27(b) shows the energy band when the photoelectric conversion device of the 22nd embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same. FIG. 28(a) shows the energy band when the photoelectric conversion device of the 23rd embodiment is in an unbiased state, and FIG. 28(b) shows the energy band when the photoelectric conversion device of the 23rd embodiment is in a biased state to perform a carrier multiplication operation. The energy band is the same. FIG. 29 is a schematic vertical cross-sectional structural diagram showing a twenty-fifth embodiment of the photoelectric conversion device of the present invention. Figure 30(a) shows the energy output when the photoelectric conversion device of the 25th embodiment is in a non-biased state, and Figure 30(b) shows the energy output when the photoelectric conversion device of the 25th embodiment is in a biased state for carrier multiplication. It is an energy empire. FIG. 31(a) shows the energy output when the photoelectric conversion device of the 26th embodiment is in a non-biased state, and FIG. 31(b) shows the energy output when the photoelectric conversion device of the 26th embodiment is in a biased state for carrier multiplication. It is an energy empire. FIG. 32 is a schematic vertical cross-sectional structural diagram showing a twenty-eighth embodiment of the photoelectric conversion device of the present invention. FIG. 33(a) shows the energy consumption when the photoelectric conversion device of the 28th embodiment is in a non-biased state, and FIG. 33(b) shows the energy consumption when the photoelectric conversion device of the 28th embodiment is in a biased state for carrier multiplication. It is an energy empire. FIG. 34(a) shows the energy output when the photoelectric conversion device of the 29th embodiment is in a non-biased state, and FIG. 34(b) shows the energy output when the photoelectric conversion device of the 29th embodiment is in a biased state to perform carrier multiplication operation. It is an energy empire. FIG. 35 is a schematic vertical cross-sectional structural diagram showing a 31st embodiment of the photoelectric conversion device of the present invention. FIG. 36(a) shows the energy consumption when the photoelectric conversion device of the 31st embodiment is in a non-biased state, and FIG. 36(b) shows the energy consumption when a bias is applied to perform carrier multiplication operation. It is an energy empire. FIG. 37(a) shows the energy consumption when the photoelectric conversion device of the 32nd embodiment is in a non-biased state, and FIG. 37(b) shows the energy consumption when the photoelectric conversion device of the 32nd embodiment is in a biased state to perform a carrier multiplication operation. It is an energy empire. FIG. 38 is a schematic vertical cross-sectional structural diagram showing a photoelectric conversion section of the 34th embodiment of the photoelectric conversion device of the present invention. FIG. 39(a) shows the energy consumption when the charging conversion section of the 34th embodiment is in a non-biased state, and FIG. 39(b) shows the energy consumption when a bias is applied to perform carrier multiplication operation. It is an energy empire. FIG. 40 is a schematic vertical cross-sectional structural diagram showing a photoelectric conversion section of the 35th embodiment of the photoelectric conversion device of the present invention. FIG. 41 is a schematic sectional view of the vicinity of the light receiving section of an embodiment of the photoelectric conversion device of the present invention. FIG. 42 is a longitudinal sectional view showing the structure of a conventional APD for optical communication. FIG. 43(a) is a vertical cross-sectional view of a conventional APD for optical communication with a step-back structure, FIG. 43(b) is a structural diagram of the energy band of the bandgap gradient layer when no bias is applied, and FIG. 43(c) ) is a structural diagram of the energy band when a reverse bias voltage is applied. 101-n+ type InP layer, 102-n type InGaAs layer, 103-... n-type InP layer, 104-p'' type InP
Layer, 105... Window, 106... P electrode, 107...
・N electrode, 108... Passivation film, 201.203, 205, 207, 209... Step back structure layer, 211... P layer, 215... n
゛Layer, 213...electrode, 214...electrode, 310...
...Light absorption layer, 301, 303, 305307.309
... Step-back structure layer, 311... P-type semiconductor layer, 315... N-type semiconductor layer, 313314... Electrode, 316... Glass substrate, 401... Cr electrode,
402... Charge injection blocking layer (n-type a-3z+-xGe
x:H), 403...multiplication region (a-3it-xG
ex:H"'+ a-3x+-yCy:H), 40
4...Light absorption layer (a-Si:H), 405...
・Charge injection blocking layer (p-type a-Si:H), 406-
Transparent electrode, 411, 412.413... composition change layer, 501- n-type a-3i+-xGex:HN, 502-
・Multiplication region, 503-・a-Si+H layer, 504 ・
p-type a-3i: H layer, 511.512.513.・・・
・, composition change layer (a-3i+-xGex: H-a-3i
t-, Cy:H), 515.516,517...
Step-back structure layer, 514... Layer closest to the n-type conductive layer, 601 =-n-type a-3it-y (:y:H
Layer, 602-... Multiplication region, 611, 612, 613...
... Composition change layer, 603-a-5i: H layer, 604
-p type a-3it-yCy: H layer, 701...n type silicon substrate, 702...0~layers, 703...p base region, 704...n'' emitter region, 705...
- Oxide film, 706... Gate electrode, 707... Insulating layer, 708... Emitter electrode, 708'... Base electrode, 709... Insulating layer, 711... Electrode, 712
-= n-type a-5it-xGex: H layer, 713-...
Multiplication region, 714--Light absorption layer, 715-"p type a
-3i: H layer, 716... Transparent electrode, 717... Collector electrode, 721722.723... Composition change layer,
731...Bipolar transistor, 732...P
Channel MOS transistor, 733... Capacitor, 734... Photoelectric conversion device, 735... Terminal, 7
36...Terminal, 737...Sensor electrode, 738...
・Emitter electrode, 739...Collector electrode, 740・
...One pixel cell, 741...Collector electrode, 742.
...Sensor electrode, 743°743', 743-...Drive wiring, 744...Vertical shift register (VSR),
746, 746', 746''...Vertical wiring, 747,
747', 747-... Reset switch, 750,
750', 750''...readout switch, 748.
・Terminal, 7490, ・Terminal, 751,751', 7
51″...Wiring, 752...Horizontal shift register (
H3R), 753. =Horizontal readout wiring, 754...
Reset switch, 755... terminal, 756... terminal, 757... amplifier, 758... terminal. 801... Photoelectric conversion unit, 802... Storage means, 80
3... Scanning means, 804... Reading means, 805
...Signal output section. 901...Cr electrode, 902...Charge injection blocking layer (
n-type a-3Lt-xGex:H), 903...multiplication region (a-3x+-xGex:H~a-3i+-yCy
:H), 904...Light absorption layer (a-3it-yC
y:H), 905... Charge injection blocking layer (p type a-
3il-y(y:H), 906...transparent electrode,
911.912,913...composition change layer, 1001
=-n type a-Sit-xGe, :H layer, 1002-...
Multiplication region, 1003-a-3i+-yCy:H layer, 10
04'''p-type layer 3it-yCy:H layer, 1011,1
012, 1013, -..., composition change layer (a-3i+-
, Gex:H~a-3t+-yCy:H), 1015
,1016.1017...Step-back structure layer, 1
014...Layer closest to the n-type conductive layer, 1101-n-type a-SI I-yCy:H layer, 110
2-... Multiplication area, 1111, 1112.1113...
・Composition change layer, 1103-a-3i,-,Cy:H layer, 1104-p-type a-3it-yCy:H layer, 1201...Cr electrode, 1202...charge injection blocking layer (n-type a -3L+-++Ge++:H), 12
03-Multiplication region (a-5it-XGeX:I(~a-3
it-VCV:)l), 1204...Light absorption layer (a-3it-xGex:t(), 1205 ・=
Charge injection blocking layer (p-type a-3i+-xGex:H),
1206...Transparent electrode, 1211, 1212.12
13... Composition change layer, 1301-n type a-S1+-xGex:H layer, 1302
...multiplication region, 1303 ・-a-5i, -, Ge
x: H layer, 1304 ・p type a-5il-xGex:
H layer, 1311, 13121313,..., composition change layer (a-3i+-xGex:H~a-3i1-, Cy
:H), 1315, 1316, 1317
... step-back structure layer, 1314 ... layer closest to the n-type conductive layer, 1401 ... n-type a-Sil-yCy:H layer,
1402...multiplication area! 411,1412.14
13... composition change layer, 1403-a-3i+-++G
ex: H layer, 1404-...p type a-3L+-xGex
:H layer, 1501... Cr electrode, 1502... charge injection blocking layer (n-type a-Sit-xGex:H), l 503
... Multiplication region (a-3t+-xGex: H~a-3i
+-yCy:H), 1504...Light absorption layer (a
-Si:H-a-si+-ycy:H), 150
5... Charge injection blocking layer Cr type a-3i:H), 15
06...Transparent electrode, 151L 1512, 1513
...composition change layer, 1601-=n-type a-3i+-xGex:H layer, 16
02...multiplication region, 1603-a-Si:H~a
-3i+-yCy: t (layer, 1604"・I) type a-
Si + H layer, 1611, 1612.1613. ...
, composition change layer (a-Sit-xGex: H~a-3it
-yCy:H), 1615. 1616. 16
17... Step-back structure layer, 1614... Layer closest to the n-type conductive layer, 1701--- n-type a-3l + -yCy:H layer, 1702... Multiplication region, 1711, 1712 .1
713... composition change layer, 1703-a-3i:)I-
a-3it-yCy: H layer, 1704-p type a-3i
:H layer, 1801...Cr electrode, 1802...Charge injection blocking layer (n-type a-3t+-xGex:H), 1803
... Multiplication region (a-3i+-+tGex: H-a-
3it-yCy:H), 1804-light absorption layer (a
-Si+H~a-3it-, Ge, :H), 180
5... Charge injection blocking layer (p-type a-3i:H),
1806 ・-Transparent electrode, 1811, 1812.181
3... Composition change layer, 1901- n-type a-3t+-xGex:H layer, 190
2... Multiplication region, 1903-a-3i: H-a-3
it-xGex: H layer, 1904 ・I) type a-3i:
H layer, 1911, 1912.1913. ..., composition change layer (a-3i+-xGex:H~a-5it-yCy
:H), 1915. 1916. 1917...
- Step-back structure layer, 1914... Layer closest to the n-type conductive layer, 2001 =-n-type a-3i l -yCy: H layer,
2002... Multiplication area, 2011, 2012.20
13... Composition change layer, 2003-a-3i:H~a-
3it-xGex: H layer, 2004-=p type a-3i
:H layer, 2101...Cr electrode, 2102...charge injection blocking layer (n-type a-3t+-xGex:H), 2103.
= Multiplication region (a-3it-xGex: H~a-3i+-
yCy:H), 2104...Cr electrode, 21
05...Light absorption layer (a-Si:H), 2106
-1・Charge injection blocking layer (p-type a-3i:H), 21
07...Transparent electrode, 2111, 2112.2113.
... Composition change layer, 2201-n type a-3t 1-xGex: H layer, 2
2028. . Multiplication region, 2203-=Cr electrode, 2204-a-
5i: H layer, 2205 ・P type a-Si+H layer, 221
1°2212.2213. ..., composition change layer (a-
S11-xGex:H" a-3x+-, C, :H)
, 2215 + 2216. 2217...Step-back structure layer, 2214...
Layer closest to the n-type conductive layer, 2301 =-n-type a-3i l-, C,:H layer, 23
02... Multiplication area, 2311, 2312.2313.
・・Composition change layer, 2303 ・・Cr electrode, 2304−
a-3i: H layer, 2305-...p type a-3i,-, C
y: H layer, 2401... n-type silicon substrate, 2402
...n layer, 2403...p base region, 2404.
... n9 emitter region, 2405 ... oxide film, 240
6... Gate electrode, 2407... Insulating layer, 2408
...Emitter electrode, 2408'...Pace electrode, 2
409... Insulating layer, 2411... Electrode, 2412
・-n type a-3L l -xGex・H layer, 2413-
... Multiplication region, 2414-a-3iGe, :H, 241
5-light absorption layer, 2416-a-Si:H layer, 2417.
...Transparent electrode, 2418...Collector electrode, 2421
, 2422.2423... Composition change layer, 2501... Cr electrode, 2502... Charge injection blocking layer (n-type a-3i+-xGex:H), 2503.
・・Multiplication region (a-3L+-xGe, :H-a-Si+
-yCy:H), 2504...n type a-3it-
xGex:H, 2505-light absorption layer (a-Si:H),
2506... Charge injection blocking layer (p-type a-3i:H),
2507...Transparent electrode, 2511, 2512゜251
3... Composition change layer, 2601-n type a-3i+-++Ge++:H layer, 26
02 ... Multiplication range, 2603- a-3i+-xG
ex: H, 2604・= a-3i: H layer, 2605
・P type a-3i: H layer, 2611.2612, 2613
．． ..., composition change layer (a-5i+-xGex:H-
a-3it-yCy:H), 2615, 2616
．． 2617...Step-back structure layer, 2614...
・Layer closest to the n conductivity type layer, 2701-n type a-5it-, C,:l (layer, 2702
- Multiplication region, 2711, 2712.2713... composition change layer, 2703- n-type a-3i:H layer, 2704-
a-3i: H layer, 2705 ・ID type a-3t+-yC
y: H layer, 2801...Cr electrode, 2802...charge injection blocking layer (n-type a-3i+-xGex:H),
2803... Multiplication region (a-St+-xGe+++
H""" a-3t+-yCy:H), 2804
...n type a-3it-xGex:H, 2805...
Light absorption layer (a-St+-xGex:H), 2806
... Charge injection blocking layer (p-type a-3it-++Gex
:H), 2807---Transparent electrode, 2811.28
12.2813...composition change layer, 2901--- n
Type a-3it, ++Gex:H layer, 2902-...multiplication region, 2903...a-3i+-xGex:H, 2
904...a-Si+-xGex: H layer, 2905-
p-type a-SiGe, :H layer, 2911, 2912, 29
13. -... composition change layer (a-3i+-xGex:H~
a-Si+-y(:y:H), 2915.2916.2
917...Step-back structure layer, 2914...n
Layer closest to type conductive layer, double region, 3311, 3312
．． 3313... composition change layer, 3303- n type a-3
L + -yCy: HNI, 3304-a-3i
t-yC,: 8 layers, 3305 ・p-type a-5it-yC
y: 8 layers, 3401... Cr electrode, 3402... charge injection blocking layer (n-type a-3it-*Gex:H), 3403.
・・Multiplication region (a-3i+-xGex: H-a-Si+
-yCy:H), 3404-...n-type a-5t+-
xGex:H, 3405... Light absorption layer (a-3i:H
~a-Sit-yCy:H), 3406-charge injection blocking NCp type a-3i:H), 3407 = transparent electrode, 3411.3412.3413...composition change layer,
3501 =-n type a-3i +-xGex: 8 layers,
3502-9° multiplication region, 3503...a-3i+
-xGex:H2S2O8-a-Si:H~a-Si+
-yCy: H layer, 3505- p type a-3i: 8 layers, 3
511, 3512, 3513°..., composition change layer (a
-3it-xGex+H-” a-3it-yCy:H
), 3515, 3516. 3517... Step-back structure layer, 3514... Layer closest to the n-type conductive layer, 3601- n-type a-3i + -yCy: 8 layers, 3
602 ... Increase 3001-= n type a-3it-yC
y: 8 layers, 3002--multiplication area, 3011, 301
2.3013... composition change layer, 3003- n type a-
3t + -xGex IH layer, 3004 =・a-3
i+ -xGex: 8 layers, 3005-p type a-3l+
-xGe, : 8 layers, 3101... Cr electrode, 3102... charge injection blocking layer (n-type a-Si+-*Ge*:H), 3103.
・・Multiplication region (a-3it-XGeX:H~a-SL+
-WCy:H), 3104...n type a-S
i+-xGex:H, 3105...light absorption layer (a-s
i+-ycy:H), 3106...charge injection blocking layer (p-type a-3it-yCy:H), 3107-transparent electrode, 3111.3112.3113...composition change layer, 32,' 01 ・-・n type a-St+-xGex
:f(layer, 3202-...multiplication area, 3203''・a
-3x+-xGex:H, 3204-a-3i+-yC
y: 8 layers, 3205 ・p-type a-31+-yCy: 8 layers, 3211, 3212, 3213. ..., composition change layer (a-3i+-xGe, :H~a-3it-, Cy:H
), 3215.3216.3217... Step-back structure layer, 3214... Layer closest to the n-type conductive layer, 3301- n-type a-3l 1-yCy: 1 layer, 3
302 , , multiplication area, 3611, 3612.361
3... Composition change layer, 3603 ・n type a-Sit-y
Cy: 8 layers, 3604-=a-3i: H-a-3i
+-yCy: 8 layers, 3605-p-type a-3i: 8 layers, 3701...Cr electrode, 3702...charge injection blocking layer (n-type a-3i+-xGex:H), 3703-
Multiplication region (a-3t+-xGex:H"" a-3i
+-yCy:H), 3704...n-type a-
3i+-xGex:H, 3705-light absorption N(a-3i
:H-a-Si+-xGex:H), 3706...
・Charge injection blocking layer (p-type a-3i:H), 3707
- Transparent electrode, 3711, 3712.3713... composition change layer, 3801... - n-type a-3it-xGe++: 8 layers,
3802-...multiplication area, 3803-a-3it-x
Gex:H, 3804---a-3i:H~a-3i+
-++Gex: H layer, 3805 ・p type a-3i: 8
layer, 3811, 3812, 3813°..., composition change layer (a-31+-xGex:H~a-3l+-yCy:
H), 3815, 3816. 3817... Step-back structure layer, 3814... Layer closest to the n-type conductive layer, 3901- n-type a-3i1-yCy: 1 layer, 390
2... Multiplication area, 3911, 3912.3913.
... composition change layer, 3903- n-type a-3it-mGe
m:Hj! F, 39o4-a-3i: H-a-3it
, xGe,: 8 layers, 3905 ・p-type a-3i: 8 layers, 4001-An electrode, 4002- n-type C-3i substrate,
4003 =-a-3i+-xGex:H-a-3it
-yCy:H14004-・-a-Si:H14005
-= p-type a-3i: H14゜06-...transparent electrode, 4
011,4012.4013... composition change layer, 4101-n type C-3i substrate, 4102... multiplication region, 4103- a-3i: 8 layers, 4104-p type a-
3i: 8 layers, 4111, 4112, 4113°...,
Composition change layer (a-Sit-++Gex: H~a-3it
-, Cy:H), 4115, 4116.4117...
Step-back structure layer, 4114... layer closest to the n-type conductive layer, 4201, Af2 electrode, 4202, n-type poly S1 layer,
4211, 4212.4213... composition change layer, 42
03- Multiplication region, 4204- a-Si:H layer, 4
205- p-type a-Si: 8 layers, 4206--transparent electrode, 4301--p-type silicon substrate, 4302--p layer, 4303-n base region, 4304--p'' emitter region, 4305 ... Oxide film, 4306... Gate electrode, 4307... Insulating layer, 4308... Emitter electrode, 4309... Insulating layer, 4310... Multiplication region, 431 ], --... Light absorption layer, 4312-p type a-3
i: 8 layers, 4313...transparent electrode, 4314...collector electrode, 4321, 4322.4323...composition change layer.

Claims (6)

[Claims] (1) A light absorption layer that absorbs light and has a forbidden band width Eg1, a minimum forbidden band width Eg2 that multiplies carriers generated by absorbing light, and a maximum forbidden band width Eg3 that are continuous. A photoelectric conversion device characterized in that a multiplication layer formed by laminating one or more layers of a modified step-back structure is laminated so as to be sandwiched between charge injection blocking layers. (2) The photoelectric conversion device according to claim 1, wherein at least one of the charge injection blocking layers is an n-conductivity type layer. (3) The photoelectric conversion device according to claim 1, wherein at least one of the charge injection blocking layers is made of a p-conductivity type layer. (4) The photoelectric conversion device according to claim 1, wherein at least one of the charge injection blocking layers is made of a metal that forms a Schottky junction with an adjacent semiconductor layer. (5) The photoelectric conversion device according to claim 1, wherein the light absorption layer, the multiplication layer, and the charge injection blocking layer are made of a non-single crystal containing at least Si atoms. (6) A light absorption layer that absorbs light and has a forbidden band width Eg1, a minimum forbidden band width Eg2 that multiplies carriers generated by absorbing light, and a maximum forbidden band width Eg3 that are continuous. A plurality of photoelectric conversion sections each having a multiplier layer formed by stacking one or more layers of a changed step-back structure and stacked so as to be sandwiched between charge injection blocking layers, the plurality of photoelectric conversion sections storage means for accumulating electrical signals generated by the plurality of photoelectric conversion units, scanning means for scanning the electrical signals generated by the plurality of photoelectric conversion units, and readout unit for reading out the electrical signals generated from the plurality of photoelectric conversion units. A photoelectric conversion device characterized in that a signal output section having at least one means and the plurality of photoelectric conversion sections are electrically connected.