JPS59154079A - Photoelectric conversion semiconductor device - Google Patents

Abstract

PURPOSE:To reduce the area necessary to connect elements by a method wherein the first and second electrodes are electrically connected, and an oxide insulator and a conductive oxide are provided on the side surface of a non single crystal semiconductor isolated by an open groove between the titled elements. CONSTITUTION:The third open groove 20 cuts and isolates the second electrode 4 of the first and second electrodes 31 and 11. The first electrode 2 of the first element 31 is electrically connected on the side surface 8 of said electrode 2 by a connector 30 extending along on a passivation film 33 and that 34 of the oxide insulation such as Si oxide from the second electrode 38 at a connecting path 12, resulting in the series connection of the two elements. Further, a semiconductor 3, a semiconductor having a P-I-N junction composed of e.g. an SixC1-x (0<x<1) P type-I type-microcrystallized N type Si 44, is provided.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a photoelectric conversion element ( A device in which multiple elements (also simply called elements) are electrically connected in series.
The present invention relates to a photoelectric conversion device capable of generating high voltage.

This invention uses a laser scribing method (hereinafter referred to as LS), which is a maskless process, in order to reduce the area required for connecting multiple elements to 1/10 to 1/100 of the conventional mask alignment method. It is characterized by
The arrangement, size, and shape of elements in the device of the present invention are determined by design specifications.

However, in order to simplify the content of the present invention, in the following detailed description, the first electrode on the lower side (substrate side) of the first element and the second electrode of the second element disposed on the right side thereof will be described. electrode (
The description will be based on a pattern in the case of direct electrical connection to the semiconductor (that is, the side away from the substrate).

In such a pattern, an opening groove separating the non-single crystal semiconductors of the first element and the second element is provided over the transparent conductive film (hereinafter referred to as CTF) of the first element, and is further provided over the transparent conductive film (hereinafter referred to as CTF) of the first element. The opening groove separating the two electrodes is
The feature is that it provides redundancy to improve molding yield by providing it over the semiconductor layer of the device.

According to the invention, the electrical connection at the connection part is made by connecting the first element to the first element.
The second electrode component of the second element is provided as a connecting conductor (hereinafter referred to as a connector) in close contact with the surface or side surface of the transparent conductive film constituting the electrode.

At this time, this connector uses a conductive oxide (hereinafter referred to as CO) to improve the reliability of the device.

Further, an oxide insulator of the semiconductor, typically silicon oxide, is provided between the CO and the side surface of the non-single crystal semiconductor.

That is, the present invention provides a second groove in the semiconductor to separate the first element region and the second element region, by performing laser scribing (simply referred to as LS) in an oxygen or oxide atmosphere. The surface of the semiconductor is heated to 50% by the high temperature of light.
Oxidized to a thickness of 0 to 3000^ and provided as a silicon oxide insulator in semiconductors, especially non-single crystal semiconductors whose main component is silicon, and simultaneously performs element division and padding of the semiconductor surface. It is.

Furthermore, in the present invention, this badge hasion film is made of pure S.
It is not iO□, but is a lower grade silicon oxide containing a large amount of silicon, and has porosity, and has the disadvantage of being non-uniform in thickness.

For this reason, if the connector that is installed closely on the top surface is made of metal, such as aluminum, it is necessary to prevent this metal from migrating (field diffusion) and reaching the semiconductor, causing leakage. In the present invention, the conductor itself is made of a conductor (CO) which is 11% less than 100%, to prevent the metal components of the conductor itself from migrating through the padding film and into the semiconductor. be.

In order to achieve high reliability at this connection part, the present invention provides a padding film made of silicon oxide (G) on the side surface of the semiconductor, and on this silicon oxide, a conductor as a connector is made of an oxide such as ITOl (indium tin). It is characterized by being provided with a conductor. In the conventional mask alignment method, the connecting portion required a toughness of 5 to 1 mm, but the present invention has a toughness of 350 to 307, preferably 200 to 1/10, which is 1/10 to 1/100 of that. This connecting part is 1
In a hybrid system that requires 0 to 50 stages, the area for photovoltaic generation (referred to as effective area) of all panels used as a photoelectric conversion device is 75
It is characterized by increasing the conversion efficiency from ~50% to 97% to 90%, and substantially improving the conversion efficiency by 10 to 20%.

This invention is a maskless process using the LS method,
In this manufacturing process, the opening formed in the previous process is
The image was enlarged ~300 times and projected on a television, etc., and the monitored open groove was stored in a computer (microcomputer).

Furthermore, using this imputed information as a reference, the amount of shift from there and the information stored in the memory are combined,
Define the position of the open groove created in this process.

Then, at this specified position, a laser beam for LS, for example, a YAG laser with a wavelength of 1.0 mm (focal length: 40 mm,
Laser beam diameter 2)) is irradiated.

Furthermore, it is moved at a speed of, for example, 5 m/min to create and set open grooves that are subordinate to the previous process.

By combining Kakunodo and <LS with a microcomputer, experimentally the following 2. ν
Open grooves in the next process can be created with the following accuracy.

That is, the LS of the present invention has another feature that it is an ultra-high precision system that can perform a substantially computer-controlled self-line method.

For this reason, all causes of price increases and yield decreases in manufacturing, such as mask misalignment, warping, and alignment of 41+i degrees required by conventional mask alignment methods, are eliminated at once. Features:

Conventionally, many examples of photoelectric conversion devices (hereinafter simply referred to as devices) have been known, in which a plurality of elements are arranged on the same substrate, and the devices are integrated, arrayed, or composited.

For example, JP-A-55-4994, JP-A-5542427
4 Furthermore, patent application No. 54-90097 filed by the present inventor
/90098/90099 (filed on July 16, 1972) is known.

For example, the patent application filed by the present inventor describes the semiconductor 5ix
C, -e -Si heterojunction, which is different from the case where only other amorphous S1 is used, and furthermore, as this semiconductor, in addition to the amorphous structure, hydrogen or halogen elements containing a microcrystalline structure are added. PN or P
It has the advantage of being a non-precision semiconductor with a small number of IN junctions (or one junction) in a stacked, hybrid, or hybrid format.

However, in these conventional inventions, although a vertical cross-sectional view thereof is not shown in FIG. 1, all of them employ a mask alignment method, which results in poor alignment accuracy and requires a large area for the connecting portion.

For example, when a metal mask is used, in a method of directly and selectively producing a conductive layer or a semiconductor layer, the mask that provides this selectivity may shift by 0.5 to 3 mm during film formation.

Furthermore, since the @ film component is formed on this mask, the mask is contaminated, the peripheral edge of the film formed along the mask is not clear, and there is crosstalk (leakage) between adjacent electrodes. It had many drawbacks, such as being a factor in the generation of electric current.

Furthermore, in conventionally known screen printing methods, conductors or semiconductors formed entirely on a substrate are independently and selectively etched away using a mask.

However, in this method, the process such as the process of aligning the mask for screen printing, the process of coating the resist 1-, the beta solidification process, the process of etching the conductor or semiconductor, and the process of removing the resist 1- is very time-consuming, and therefore the manufacturing process is difficult. Price increases were inevitable.

However, in the photoelectric conversion device of the present invention, particularly in the thin film type photoelectric conversion device, the conductive layer for electrode, which is each M film layer,
It has also been found that since each semiconductor layer has a thickness of 500, it is possible to manufacture the semiconductor layer without requiring automatic mask alignment using computer control.

As a result, since the mask of the present invention is not used at all in place of the conventional mask alignment process, it is a maskless process, which is extremely simple and highly accurate, and the manufacturing cost of the device is low.
Therefore, it is possible to manufacture for 500 yen/W, and the f! The purpose of the present invention is to provide an extremely innovative photoelectric conversion device that can be produced for 100 to 200 yen/W by expanding the scale of construction.

A conventional example and the structure of the present invention will be described below according to the drawings.

FIG. 1 is a longitudinal sectional view of a conventionally known photoelectric conversion device with a mask assembly.

In the drawings, a transparent conductive film (abbreviated as CTF) constituting a first electrode is selectively formed on a transparent substrate (for example, a glass plate) (1) by a first mask alignment step.

Further, semiconductor N(3) is similarly selectively formed by a second mask alignment step.

Furthermore, the second electrode (4) is formed by a third mask alignment process.
is provided.

In FIG. 1, there is a connecting part (12) between the elements (11) and (31>), and in the connecting part, the semiconductor layer (3) covers one side (16) of the CTF, and the other side Since the surface (14) of CTF is 1° so as not to be covered by the semiconductor layer (3), the distance between CrF2 (13) is 1 to 5++on, for example, 3+n.
m's.

Requires a gap.

Furthermore, the first electrode (37) and the second electrode (38) are (1
It is electrically connected on the surface of 4), but this part is connected to (39)
1 to 5 mm, for example, 31 mm, because the second electrode of the
Especially this second electrode (39) characterized by a gap (6)
In order to prevent short-circuiting with the first electrode (37), the alignment accuracy at (28) is extremely important for manufacturing yield, and as a result, the connecting portion (12) has become wide.

In addition, the first electrode (37) and the second electrode (39) do not leak through the semiconductor surface (28) (resulting in a decrease in reliability).

Therefore, the first electrode (37) and the second electrode (39) can be separated without adding any steps in the manufacturing process.
There has been a strong demand for a structure in which the surface of the semiconductor between the two is covered with a banshihation film, which is extremely important not only for improving manufacturing yield but also for high reliability.

The present invention meets this objective.

Also, if the gap of this connection part is 3nll11, for example, 20c
11Iy60cm to November 5+nm (20cm
15mm), if the element end is 5mm long, then 33
It is a step connection, and the connecting part has a total of 10cI11 (2
As a result, the effective area remained at 75% considering the peripheral area.

The present invention eliminates such process complexity and reduces the effective area to 86.
In addition, the semiconductor between the upper and lower electrodes is protected by a pansivation gland without being exposed, which is an innovative photoelectric conversion device.

The details of the invention are shown below in accordance with the drawings.

FIG. 2 is a vertical W [side view] which does not show the manufacturing process of the present invention.

In the drawings, the transparent substrate (1) is a glass plate (e.g. 0.6-2.2 mm, e.g. 1.2 mm, length [
In the drawing, the horizontal direction) 60 cm, +4] 20 cm) was used.

Further, a transparent conductive layer 11 is applied to the entire upper surface, for example, ITO (approximately 1500 Å) +SnO, (200-40
071) or a transparent conductive film whose main component is tin oxide added with a halogen element (1500 to 2000 i >
was formed by a vacuum vaporization method, LPCVD method, plasma CVD method, or spray method.

After that, from the bottom or top of this substrate, a G laser processing machine (Nippon Rede) is used to produce an output of 3 to 6 W (focal length: 40 mm).
mm), with a spot diameter of 20 to 5ν typically 307
' is controlled by a microcomputer, irradiates a laser beam from above, and forms a first groove (13) for a scribe line from the scanning number, and
1) The first electrode plate (2) was prepared in (11).

The open Mj (13) formed by LS was approximately 3 mm wide and 20 cm deep to completely separate each of the first electrodes.

For this reason, as is clear from the drawing, part of the board (1) was gouged out to a depth of 300 to 1,300 people (four parts).
60)).

Thus, the first element (31) and the second element -(11
) is 15-30m+a, for example, 1
5 + n tll.

As described above, using the LS method, the CTF (2
) was cut and separated to form a first open groove.

After this, plasma CVD or LPGV D is applied to this upper surface.
The non-single crystal semiconductor layer (3) having a PN or PIN junction is formed by a method of 0.2 to 0.8. It is typically formed to a thickness of 0.2 mm. ' A typical example is a P-type semiconductor (SixC1-cxJ, Y about 1
00λ) - I-type amorphous or monohymiamorphous silicon semiconductor (approximately 0.p) - N-type finely divided crystal (5th
A non-single crystal semiconductor with one PIN junction made of a semiconductor with 20OA) or a P-type semiconductor (SiX
C+J- ■ type, N type, P type Si semiconductor - I type 5ixGc
, <Tandem-type PINPIN having two PIN junctions and one PN junction made of semiconductor-N-type Si semiconductor,
, , , PIN junction semiconductor (3).

Such a non-single crystal semiconductor (3) was formed with a uniform thickness over the entire surface.Furthermore, as shown in FIG.
A second open groove (18) was formed across the left side (first element side) of 3) by the second LSI process.

In this drawing, the first and second open grooves (13), (1
8) is shifted by 5 degrees.

The laser light was irradiated either from below the glass (1) or from above the substrate.

The second open groove (18) thus forms a side surface (8) of the first electrode.
(9) was exposed.

2. The side surface (9) of this second groove may be on the left side of the side surface (16) of the first electrode of the first element, and is shifted by 10 to 10% toward the first electrode. That is, it is characterized in that it is provided over the first electrode position of the first element.

As an extreme example, as shown in FIG. 2(B), it may enter the inside (9) of the first electrode (37).

Furthermore, in the present invention, as shown in the conventional example, it is not necessarily necessary to expose the surface (14) of the first electrode (see FIG.
The pressure was applied a little too forcefully, and the entire CTF (37) was removed in the depth direction, and as a result, the connector with the second electrode (38) was attached to the side surface (8) as shown in Figure 2 (C). There is no practical problem even if they are in close contact with each other.

In other words, it is extremely important for the industrial application of the present invention to be able to provide a margin in construction for variations in the intensity of the output pulse of the RED light and the depth of the groove.

In the present invention, the second open groove (18) is formed in an oxide atmosphere, that is, in oxygen or air.

In addition, this substrate is maintained at a temperature between room temperature and 200°C, for example 150°C, so it instantly melts when irradiated with laser light, and the instantaneous irradiation temperature reaches approximately 2000°C, so this high temperature causes the semiconductor to vaporize. It will be removed.

Furthermore, due to this remaining thermal processing, 500 to 3000 oxides of the semiconductor are formed on the side surface (32) of the semiconductor (3).
It was possible to form it to the thickness of a person.

In the present invention, the semiconductor is mainly composed of silicon, so the oxide is mainly composed of silicon oxide, and in addition, the first CT
As a result of measurement using SIMS (Ion Micro Analyzer), it was found that several percent of tin, which constitutes F(2), was contained.

Furthermore, 3000 times more than SIEM (scanning electron microscope)
When examined at 50,000 times magnification, this oxide insulator is porous and extremely uneven, and the thickness of the formed film is 50,000 times.
The number had grown to between 00 and 1000 people.

The reason for this is that in this example, the laser beam was discontinuously irradiated with pulsed light, and the irradiation rate of 307" was 5 m/min (8.3 cm).
+n 7 seconds) and a frequency of 50 KIIz.

In order to make this oxide thicker and more uniform iY, it was possible to reduce the scanning speed to 0.5 m/min.

Since the oxide insulator formed on such a semiconductor has a resistance value of 10 cm, it is not only capable of isolating the connector and the semiconductor, but is also extremely useful as banshihation for the semiconductor. It was 〃ノ.

Furthermore, even if it is discharged for 1000 hours at 150'C, the leakage at the edge of this semiconductor is more than IOA/c+n (the difference in thickness of the peripheral part is less than 108 per jam RIJ "320cm").
x 4 per 46 cm length for a 15 mm element
I was able to prove myself the difference of X108]'-.

In FIG. 2, as shown in FIG. 2<C>, a second electrode m(4) on the back side is further formed on this upper surface, and a third opening groove for glue cutting and separation is formed in the third LS. (20) was obtained.

This second electrode (4) is a conductive oxide (GO) (45)
The membrane was formed to a thickness of 700-1400 Å.

As this CO, ITOCM (insium tin oxide) (45) is formed, and aluminum to which 1% or less of silver or silicon as a reflective metal (46) is added is added to the top surface to a thickness of 300 to 3000'h. was formed.

Furthermore, it is effective to form Ninogel on the upper surface as an electrode for external connection and as an anti-oxidation agent for aluminum.

One row had a 3m structure with ITO at 1050A, silver or aluminum at 1000A, and Ninogel at 1500A.

The ITO and reflective metal are used to promote reflection of long-wavelength light on the back side and to effectively photoelectrically convert long-wavelength IQ light of 600 to 800 nm.

Furthermore, the Nisogel is used in the external extraction electrode (2
3) It is useless to improve the adhesion with 3).

These were formed at temperatures below 300'C using electron beam evaporation or CVD to avoid deteriorating the semiconductor layer.

This shows the case where the second electrode is irradiated with a laser beam from the nickel and cut W1 apart to form an open groove (20).

This laser light comes into close contact with the semiconductor, especially the top surface, and
Slightly scoop out the 0λ N or P type thin semiconductor jG (40), and remove the remaining ring or conductive semiconductor in the open groove between the adjacent first element (31) and second element (11). This prevents crosstalk (leakage current) from occurring.

In particular, this semiconductor (3) forms, for example, one I"IN junction with a P-type semiconductor layer (42), an N-type semiconductor layer (43), and an N-type semiconductor layer (44), and this N-type semiconductor layer forms a microcrystalline semiconductor layer. or has a polycrystalline structure.If the so-called electrical conductivity is as high as 1 to 200 cubic meters (Acm), the N-type semiconductor layer of the present invention is cut out, and the fourth part (40) is made into an intrinsic semiconductor. Providing a passivation layer of an oxide insulator such as silicon oxide (34) on this semiconductor to prevent the generation of leakage current is extremely effective in terms of high reliability.

It is preferable that this extrusion exceeds the N-type semiconductor layer and does not reach the CTF for the first electrode.

Thus, as shown in FIG. 2(C), a plurality of elements (
31) We were able to create a photoelectric conversion device that directly connects (11) at the connecting part.

FIG. 2(D) shows the attempt to further complete the present invention as a photoelectric conversion device, that is, as passivation I9, a silicon nitride film (21) with a thickness of 500 nm is formed by plasma vapor phase method.
Formed uniformly to a thickness of ~2000' h, leakage current 11 between each element was further prevented from occurring due to adsorption of moisture, etc.

Further, an external lead terminal (23) was provided at the peripheral portion (5).

These were filled with an organic resin (22) such as polyimide, polyamide, Kapton or epoxy.

In this way, the irradiation light (10) appears and the base 4 as in this example is
N (60cm x 20cm)
4.35+n+n 1501I in the connecting part, IIJ 101I1m in the external extraction electrode part, and 4mm in the periphery, it has 401ix within 580mm K192mm, and has an area of 192mm x 14.35mm 4
0 stage 11'02 cmL (t) 91.8%) could be obtained.

As a result, 10.6% (1,05AI
'cnF), the panel has a conversion efficiency of 6.7
% (theoretically 9.7%, but due to the resistance of the 40-stage connection, the actual conversion efficiency decreased) (8 Ml (10
At 0 mW/cn:) ), an output power of 73.8 W could be achieved.

Furthermore, when this panel was subjected to a high-temperature storage test at 150'C, no decrease of 10% or less was observed after 1000 hours, for example, 4% at worst with 20 panels, X = 1.5%.

This means that when conducting reliability tests using the conventional mask method, the number of inoperable panels decreased to 1 in 10 hours.
Considering that 7 cards were generated, it was amazing (direct).

As described above, the present invention also provides that the semiconductor surface is not directly clouded between the first electrode and the second electrode, the surface is padded with the semiconductor oxide insulator, and in addition, contact (partially It has been found that the required area is sufficiently reduced to less than 1/10 compared to the conventional method.

FIG. 3 is a longitudinal cross-sectional view and a plan view (end portion) showing the most typical positional relationship of the grooves that are formed during three LSI cycles.

The numbers and steps are the same as in FIG.

FIG. 3(A) shows the first open groove (13) and the first element (31).
), a second element (11), and a connecting portion (12).

As is clear from the drawing, the first groove (13) is located on the substrate (
1) is slightly hollowed out.

Furthermore, the second opening! (18) is provided over the first electrode (2) side of the semiconductor (3) that constitutes the first element, and both of these are removed.

Therefore, the first electrode (2) of this first element (31)
and the second electrode of the second element (11) are connected to the connecting portion (12).
Then, from this second electrode (3 days), connect the first electrode (30) to 2) side (8
), and the two elements are connected in series.

Further in the drawing, a PN or 1) semiconductor with at least one IN junction (3) here one 5ixC+
-x (0<x<l) P type - ■ type Si - microcrystalline N type Si (44) A semiconductor having one f'lN junction is provided.

This third open groove (20) is shifted to a depth of approximately 31 mm toward the first element (31).

For this reason, the right end portion of the third open groove (2o) is provided under the connector portion (30).

Thus, the first and second summation 7- (31) (Ii
) are electrically cut and separated, and the leakage between these electrodes is also reduced to IOA/cm (Icm
It was possible to reduce the size to less than H), which is on the order of a medium IOA.

FIG. 3(B) shows a plan view, and the first, second, and third open grooves (13
), (18), (20) or set number -ζ.

In order to reduce leakage in this direction, the semiconductor (3
) covers the first electrode (2), thereby reducing short-circuits between the first and second electrodes.

In addition, at the end of the element, the first electrode (2), the semiconductor, and the second electrode (4) were scribed (50) at once by LS.

In this case as well, a badge hasion film is similarly formed on the side surface of the semiconductor.

In this drawing, during the first, second, and third lectures, 50~
2-, and the connecting part is November 50-89, typically 12
I was able to put the helmet on myself.

The above YAG rate Subono l-Ji# is its output 3.
~! l+W (20, -L') 4~7W (30,
m'), or by further reducing the spot diameter based on the technical concept, the necessary area for this connection can be made smaller, and the area J (effective It has an inventive step in that it can further improve the efficiency (rate).

FIG. 4 is an enlarged view of the external lead-out electrode section when small photoelectric conversion devices, rather than large panels such as those for calculators, are to be mass-produced at the same time.

FIG. 4(A) corresponds to FIG. 2, but the external extraction electrode part (5) has a pad (49) that contacts the conductive rubber electrode (47), and this bat' (49) It is connected to the second electrode (upper electrode) (4).

At this time, even if the pressure on the electrode (47) is too strong and the band (49) penetrates the first electrode (2) and semiconductor (3) below, the band (49) and (2) should not be short-circuited. Open/l5
7 (13) are provided.

In addition, the outer part is the first electrode, the semiconductor, and the second electrode 4 (6).
) is separated.

Furthermore, in FIG. 4(B), another pad (48) connected to the lower first electrode (2) is made of a second electrode material (18).
They are connected to each other.

The rough pad (48) is in contact with a conductive rubber electrode (46) and is electrically connected to the outside.

Here too, pad 2) (48) is completely electrically isolated from the adjacent photoelectric conversion device by the open groove (1'8) (20) (50), and the glass is cut between these devices in a post-process. By doing so, a large number of photoelectric conversion devices can be produced in one panel without using any alignment masks.

For example, 6cm AF on a 20cm x 60cm panel.
If we try to make a 1.5 ctn photoelectric conversion device (for calculators), we can see that 130 solar cells for calculators can be made at one time.

In other words, the photoelectric conversion device is kept in an organic resin mold (22) except for the electrode part (5) (4'5), and then when making a small power solar cell, the glass is cut into IJJW.
All you have to do is i.

In addition, each of these panels, for example, 40 cτn x 4 o or 60 cm x 20 cm, is assembled in series within an aluminum sash frame to be paso gauged.
It is possible to provide a 120 cm x 40 cm NEDO standard high power panel.

In addition, in this NEDO standard panel, a fluorine-based protective film is attached to the reflective surface side (upper side in the drawing) of the photoelectric conversion device of the present invention using Schiefle, so as to increase mechanical strength against wind pressure, rain, etc. is also valid.

In the present invention, glass is particularly used as the substrate among light-transmitting insulating substrates.

However, it is effective to use, as this substrate, a flexible organic resin or a composite substrate in which silicon oxide or silicon nitride is formed on an organic resin to a thickness of 0.1 to .

In particular, when this composite substrate is applied to the above embodiment, silicon oxide or silicon nitride damages the CTF on the upper surface.
The so-called blocking effect, which prevents the formation of a mixture of substrate and CTF, was particularly effective.

Furthermore, the details of the present invention will be supplemented by describing examples below.

Example 1 This embodiment is illustrated according to the drawing in FIG.

That is, chemically strengthened glass with a thickness of 1.1 as the transparent substrate (1)
mm, width 60 cm, and 11 J 20 cm were used.

A silicon oxide film was applied to the top surface to a thickness of 0.1 to form a blocking layer.

Furthermore, CTF on top of that is 11'01600^+S￥03
007Y was produced by electron beam evaporation.

Thereafter, a first groove was formed using a YAG laser with a spot diameter of 3 [z4 and an output of 41 mm] controlled by a microcomputer at a scanning speed of 5 m/min.

In order to completely cut the CTF with this output, a V-shaped groove of about 11" in diameter and about 3000 M in depth was created in the center of the open groove by melting away the glass substrate.

The element area (31) (11) was set to 15 mmrlJ.

Thereafter, a non-single crystal semiconductor having one PIN junction as shown in FIG. 2 was manufactured by a known PCvD method.

Its total thickness was approximately O0-.

After that, monitor the first opening groove on the TV, and shift the first element (3I) from there to Shift 1 to Suboso 1.
- A second open groove (18) was prepared as shown in FIG. 2(B) by LS at a temperature of 5 cm in the atmosphere for 100 hours with diameters so and a'.

Thus, on the side surface (32) of the semiconductor (3), silicon oxide (33) for perforation was formed to a thickness of about 100 ml.

A pole (45) connector (30) was configured.

Further, a third opening t@(20) is similarly formed in an oxidizing atmosphere by LS at a depth of 50/lI from the second opening groove (18) by shifting it to the first element (31) side. Figure 2(C) was shown in iS.

The laser beam was used at the output section, and the other conditions were the same as those for producing the second open groove.

In this way, FIG. 2(C) was produced.

After the process shown in FIG. 2(C), the first electrode, the semiconductor, and the second electrode are all rectangularly scanned with a laser beam output of 6-4 mm from the edge of the glass to form a rectangular shape on the edge of the panel. Prevents electrical short circuit with the frame.

After this, Pudgehesion D'A (21) was PCVD
A silicon nitride film with a thickness of 1000λ was fabricated at a temperature of 200°C by the method.

Then tOcm 'X 60cm (7) Hanelni 1
We were able to fabricate 40 stages of 51I1m elements.

The effective efficiency of the panel is AMI (100mW/cn
i''), we were able to obtain a power output of 6.7% and a cuff output of 3.8W.

The effective area is 1102 cm, and the total panel area is 91.8 cm.
% could be used effectively.

Example 2 Substrate glass: thickness 1.1 mm and size 20 cm k'
60 cm was used. Furthermore, a plurality of photoelectric conversion devices for calculators were fabricated on the same substrate with a length of 1.5 cm and a length of 50 m. Here, the element shape was a 5-stage continuous array with a length of 9 mm and a length of 13 mm.

The connecting part is 109fiI, and the external electrode is shown in Fig. 4 (A).
It was provided as structure (B).

As a result, he was able to make 160 calculator devices at once.

Tests were conducted with a fluorescent knife 200 x for an effective conversion efficiency of 4.5%.

As a result, a final manufacturing yield of 83% was achieved.

This can only be achieved by 40-50% using conventional methods,
Moreover, considering that the required area of the connecting portion was large and the effective conversion efficiency could only be obtained up to 3.2%, it was extremely effective.

The rest is the same as in Example 1.

Example 3 This example is Example 2, and the substrate was made of polyimide resin, which is a light-transmitting organic resin, and had a thickness of 150/14.

Furthermore, as a blocking layer, o, :P silicon oxide is exploded at a temperature of 250C by the reaction of silane and carbon dioxide gas by plasma gas phase method,
Solo soaking jml was used to prevent injuries caused by S.

The strain is the same as in Example 2.

In this method, the cost of the substrate was 30 yen in Example 2, but this could be reduced to 2r'l/calculator element.

In addition, since each calculator element can be separated from the sheet by cutting or using scissors, it is extremely easy to process and is inexpensive.

Furthermore, when cutting from this sea 1-, automatic moon 3) r became possible by LS using strong pulsed light of 10 to 15 W.

In this example, as shown in FIG. 2(D), by overlapping the upper protective organic resin (22),
It has a structure in which a photoelectric conversion device is sandwiched between organic resin sheets, has flexibility, and can be mass-produced at extremely low cost.

The yield in this example was 160 games, of which 72%.
with an effective conversion efficiency of 4.5% as the lower limit? I was able to do it.

In FIGS. 2 to 4, light was incident from the lower translucent insulating substrate.

However, the present invention is not limited to the light incident side.

[Brief explanation of drawings]

FIG. 1 is a longitudinal sectional view of a conventional photoelectric conversion device. FIG. 2 is a longitudinal sectional view showing the manufacturing process of the photoelectric conversion device of the present invention. FIG. 3 is a longitudinal sectional view of the photoelectric conversion device of the present invention. FIG. 4 is a partially enlarged vertical sectional view of another photoelectric conversion device of the present invention. Patent applicant: Semiconductor Energy Research Institute Co., Ltd. Representative: Shun Hei Yamazaki? I tp tt [ ” 1 1υ house 1■ Japanese cabbage 30 (A) 18su CB) Suspension 4■

Claims (1)

[Scope of Claims] l. A first electrode of a transparent conductive film on an insulating substrate; a non-single crystal semiconductor that is closely placed on the electrode and capable of generating optical electric power by irradiation with light; A photoelectric conversion device in which a plurality of photoelectric conversion elements each having a second electrode in close contact with each other are electrically connected in series and disposed on the insulating substrate,
The first electrode is electrically connected to the second electrode of the second element adjacent to the first photoelectric conversion element, and the non-uniform electrode is separated by the groove of the first and second photoelectric conversion elements. A photoelectric conversion semiconductor device, wherein a side surface of a crystalline semiconductor is provided with an oxide insulator of the non-single crystal semiconductor, and a conductive oxide on the oxide insulator. 2. The photoelectric conversion semiconductor device according to claim 1, wherein the oxide insulator adjacent to the opening provided by the laser scribing method contains silicon oxide as a main component.