JPS6331951B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a hybrid photoelectric conversion device in which a plurality of photoelectric conversion cells are arranged on a transparent substrate.

This invention relates to combining photoelectric conversion cells (hereinafter simply referred to as cells) on a substrate, and aims to increase the visual value of the entire device by reducing the distance between adjacent cells to 300μ or less, which is difficult to distinguish with the naked eye. The purpose is

Therefore, in the present invention, a first electrode on a transparent substrate in a cell provided in an active region, a non-single crystal semiconductor on this electrode that generates a photovoltaic force by light irradiation, and a first electrode on the semiconductor. By making the two electrodes have approximately the same shape and approximately the same arrangement (self-registration) structure, it is possible to avoid a decrease in manufacturing yield due to deviations in the alignment accuracy of composite, and also to avoid this self-registration (hereinafter referred to as SG) structure. ) using a scribing method using laser light, the distance between each cell is 300μ or less (0.3mm or less)
Preferably, the thickness could be set to 30 to 150μ. That is, after the first electrode, semiconductor, and second electrode have been formed, laser light is irradiated from the transparent substrate side to instantaneously heat and vaporize all of them.

For this reason, it is not necessary to match each mask with high precision, which is necessary for compounding, and the feature is that all the masks are finally formed simultaneously using the SG method.

Conventionally, PIN junction heterojunction or PINPIN...PIN junction and multiple
They attempted to generate photovoltaic force through light irradiation using a bonding method in which PIN and PN junctions were stacked.
However, since the upper and lower electrodes of a semiconductor having such a junction are connected in series, the lower electrode of one cell must be electrically connected to the upper electrode of the adjacent cell, and each cell is electrically connected to each other. The necessary condition was that the system be isolated from the outside.

FIG. 1 shows a typical example of a conventional structure.

FIG. 1A is a plan view of the photoelectric conversion device 30 viewed from the back with the transparent substrate 2 facing downward. In the drawing, active region 1 that generates photovoltaic force by light irradiation
0 and a non-active region 11 having a connecting portion 18 connecting each cell 1, 1'. A-A′ in Figure 1,
Figures 1B and C correspond to the vertical cross-sectional views of B-B'.
It is shown in As is clear from the correspondence between A, B, and C, in the conventional example, in the active region, each cell 1, 1' has a transparent conductive film CTF of the first electrode on the glass substrate 2. are separated from each other. Further, the semiconductors 4 are connected to each other.
Further, in the non-active region, the upper electrode of the cell 1 is connected to the lower electrode of the cell 1 by a connecting portion 18, and this is repeated to connect five cells in series between the external electrodes 8 and 9. The number and size of these cells are determined by design specifications.

However, at first glance, this conventional structure appears to have a high manufacturing yield because the semiconductor 4 is the substrate. However, in reality, three types of masks are used, and if there is even a slight deviation between the first mask and the third mask (i.e., 1 to 3 mm in the case of a metal mask).
(It is only natural that the difference in the angle of view is different) A vertical cross-sectional view as shown in FIG. 1D is created. As a result, in B,
It was found that what had 12 cells and 13 isolation areas became D's 15 cells and 14 isolations, effectively reducing the cell area by 20 to 40%. . Furthermore, since a mask is used, the isolation area of B is 1~
2mm, for example, 1.5nm, so if the cell width is 10mm and there is a deviation of 2mm, the cell width 15 will be 8mm, the isolation width 14 will be 3.5mm, and 30
The effective area decreases by nearly %.

Therefore, it has been extremely desirable to make the combination of the upper and lower electrodes self-registered in order to improve the efficiency.

Furthermore, in the conventional example shown in FIG. 1, a method is used in which a metal mask is arranged to reduce the cost, and the electrodes 3 and 5 are selectively formed only in the cell region 12. However, in such a method, when the mask is used 10 to 30 times, the same metal film is formed only in one direction of the mask, which causes stress, and the surface on which it is formed loses its closeness and starts to float. As a result, metal and CTF wrap around between the mask and the base, and as shown in FIG.
6, it was found that adjacent cells would shoot or leak in areas where electrical isolation was required. For this reason, attempts have been made to widen the distance between the electrodes, but if the distance is increased to 1 to 2 mm or more, this gap not only does not provide an effective area, but also visually reduces the commercial value.

Furthermore, in order to eliminate floating due to warping of the mask, it is possible to increase the thickness of the mask from 300 to 500 μm to 3 to 5 mm. This eliminates the warpage, but due to the thickness, when forming the electrodes 3 and 5, another drawback arises in that the ends become thin and shaded.

For these reasons, there has been a need for a photoelectric conversion device based on a completely new structure and manufacturing method that allows mask alignment at the connecting portion to be performed with low precision and substantially high precision mask alignment in the active region.

The present invention has been made in response to such a need, and details thereof will be described below with reference to the drawings.

FIG. 2 shows the manufacturing process and apparatus of the photoelectric conversion device of the present invention.

In the drawings, a light-transmitting substrate (for example, glass) is used as the substrate. This drawing shows a case where five cells are connected in series. That is, the photoelectric conversion device of the present invention has an active region 10 and an inactive region 11, and all cells in the active region have a first electrode and a non-single crystal semiconductor on the lower side, and a second electrode on the upper side. They were self-registered (SG) and had approximately the same shape and arrangement.

This is because the first electrode, the semiconductor, and the second electrode were provided all over the active region, and then all of them were scribed with a laser beam at the same time. In particular, this laser (here, a YAG laser) was used to scribe from the transparent substrate side according to a pattern stored and controlled by a microcomputer. As a result, SG became possible. Furthermore, since the laser spot is generally 30 to 50μ (3μ is possible due to its structure, but considering the yield, we used a relatively long focal length of 30μ).
Its width is 10 to 300μ, preferably 30 to 50μψ.

In FIGS. 2A, A-1, and A-2, the active region 10 and the non-active region connecting electrode 6 are connected to the electrode 3 formed by the first light-transmitting conductive film using the first mask.
was formed on the substrate 2.

This CTF is formed by laminating ITO (indium oxide containing 10% or less of tin oxide) or tin oxide in a single layer or multiple layers. Generally, it is formed to a thickness of 1500 to 2500 Å using electron beam evaporation.

In the drawing, vertical sectional views taken along lines A-A' and B-B' at A are shown corresponding to A-1 and B-1, respectively. In this drawing, the mask is only the shaded area in the non-single-crystal region 11 of A, and the pattern is simple, so the mask is inherently difficult to lift off the substrate. In addition, this mask has the feature that it does not require low alignment accuracy and there is no problem even if it is slightly lifted from the substrate 2.

Next, as shown in FIG. 2B, a non-single crystal semiconductor is formed in the active region 10. The mask at this time has only diagonal lines and is a simple pattern. Figure 2 B C
-C' and D-D' vertical cross-sectional views are shown corresponding to B-1 and B-2.

Thus, the active region has a substrate 2 as shown in B-1.
A first electrode made of CTF and a non-single crystal semiconductor 4 that generates a photovoltaic force upon irradiation with light were formed thereon.

This semiconductor 4 is, for example, a P type of SixC _1-x (0<x<1 generally x=0.7 to 0.8) with a thickness of about 100 Å.
Furthermore, we created a PIN junction structure of an I-type semiconductor mainly composed of silicon doped with hydrogen or halogen elements to a thickness of 0.4 to 0.6μ, and an N-type semiconductor mainly composed of microcrystalline silicon. . Of course, this is P(SixC _1-x x=0.7~0.8)-I(Si)-N(μCSi)
-P(SixC _1-x x=0.7~0.8)-I(SixGe _1-x x
It may be a tandem structure of PINPIN structure such as =0.6 to 0.8)-N(μCSi).

Next, the pattern shown in FIG. 2C was formed using a third mask. Vertical sectional views corresponding to lines E-E' and F-F' in FIG. 2C are shown in C-1 and C-2.
As is clear from this drawing, the electrode 6 of the lower connecting portion and the electrode 7 of the upper connecting portion are in ohmic contact to form the connecting portion 18. In this state, the active region only constitutes a single layered structure, and as is clear from the vertical cross-sectional view of C-1, only the second electrode 5 is formed on the semiconductor 4. do not have. This second electrode is made of ITO with a thickness of 900 to 1300 Å, e.g.
Provided with a thickness of 1050Å, and further coated with silicon or chromium,
A metal mainly composed of aluminum doped with titanium was formed to a thickness of 1000 to 2000 Å. Of course, if reliability is not important, ITO may be removed. Moreover, ITO alone was sufficient for this electrode.

In order to improve the characteristics by utilizing the reflected light from the back electrode, the above-mentioned ITO+A1 was preferable. Furthermore, only ITO was preferred for improved reliability. This is because the metal of the back electrode and the semiconductor tend to react.

After this, in FIG. 2B, the laser scribe 20
I did this. This was done using a YAG laser (wavelength: about 1 .mu.m) with an average output of 3 to 5 W from the glass substrate side, a beam diameter of 30 to 50, and a beam scanning speed of 1 to 10 m/min, generally 3 m/min.

Thus, G-G', H-H', I-I',
Corresponding to J-J', Fig. 2 D-1, D-2, D-
3, D-4.

As is clear from this drawing, a first electrode 3, a semiconductor 4, and a second electrode 5 are arranged on a transparent substrate 2 with a width of 10 to 10 mm.
300μ preferably 30~100μ scribe line 2
0, they are provided in approximately the same shape and in the same arrangement.

In FIGS. 2D to D-4, these upper surfaces are coated with an organic resin 22 such as silicone, epoxy, or polyimide to a thickness of 1 to 20 μm. Another feature of the present invention is that this laser scribing step D is performed from the glass side. This is to prevent the thin film-shaped first and second electrodes from shooting or leaking from each other by being heated by laser irradiation and being spattered and scribed to the outside.

By irradiating this laser light from above in the drawing, the second electrode is laser annealed, and the first electrode and the semiconductor are evaporated and diffused, causing them to be completely unusable. has been experimentally determined by the present inventor.

That is, the present invention is only possible by using a light-transmitting and somewhat heat-resistant substrate, such as a glass substrate, and by irradiating laser light from the glass substrate side. As a result, the scribe width is preferably 10 to 300μ.
The width is 30 to 100 μm, which is impossible or difficult to observe with the naked eye, and this can be applied not only to the laser 20 between each cell, but also to the scribe 21 between each photoelectric conversion device.

As a result, as is clear from this drawing, this photoelectric conversion device is, for example, 1.5 cm x 1.5 cm as shown in the drawing.
Instead of making one photovoltaic device with an active area of 5.4 cm and a non-active area of 5 mm x 5.4 cm on a glass substrate measuring 12 cm x 5.4 cm, a 20 cm x 40
It is possible to make a large number of photoelectric conversion devices at once on a large glass plate of cm or 20 cm x 60 cm or 40 cm x 120 cm. Finally, it can be seen that it is sufficient to divide these into individual photoelectric conversion devices.

Of course, a large number (100~
Even in the conventional example shown in FIG. 1, it is not impossible to produce 1,000 photoelectric conversion devices and finally divide them. However, in such a case, the conventional method has its own limitations because the mask requires a high degree of alignment accuracy, and the possibility of the mask lifting off from the substrate is extremely problematic.

Figure 3 A, B, C, and D are Figure 2 A, respectively.
The outline of large-area patterns corresponding to B, C, and D is shown.

Area 30 surrounded by a dashed line in the drawing...
30 indicates independent photoelectric conversion devices.
The active area is 10,10' and the inactive area is 11.
It is very simply arranged in the form of a strip. For this reason, since a mask is used only to connect the band-like rough (loose) adjacent cells, the accuracy of this alignment may be loose, and mass production is extremely easy.

Furthermore, as is clear from FIG. 2D, the effective area of the cell is effective except for a very small part of the active region with a width of 10 to 300 μm;
95±2% or more can be obtained, compared to the conventional example of 80±30.
%, the structure of the present invention is much superior.

From the above, the present invention has a large active area that can be manufactured at 1/3 to 1/5 of the cost compared to conventional methods, since it is only necessary to increase the area and finally divide it into each photoelectric conversion device. Due to the self-registration method, the effective efficiency of cells is high and there is little variation.The production yield is high because high mask alignment accuracy is not required.The scribe line between each cell is self-registration, and the laser beam spot diameter is Change it to 10-300μ, preferably 30-300μ, which is 1/10 to 1/50 of the conventional 1-1.5mm.
It was possible to make it 100μ. As a result, hybridization was not visible to the naked eye, and high added value was achieved.There was no blurring at the periphery of the cell due to mask lifting, and the ridged rainbow structure at the periphery of the conventional example was not seen. It has many features, such as the ability to reduce the amount of waste and give it high added value.

The above description is not limited to the patterns of FIGS. 2 and 3 of the present invention. The number and size of cells are determined by the design specifications. For semiconductors, plasma CVD method or low pressure CVD method was used.
Applications are possible not only to PIN junctions, heterojunctions, and tandem junctions, but also to many structures whose main component is non-single-crystal silicon.

[Brief explanation of the drawing]

FIG. 1 is a vertical sectional view of a conventional photoelectric conversion device. FIGS. 2 and 3 show a plan view and a vertical sectional view of the photoelectric conversion device of the present invention according to the manufacturing process.

Claims (1)

[Claims] 1. An active region that generates a photovoltaic force by light irradiation on a transparent substrate, and a connecting portion and external electrodes that connect a plurality of photoelectric conversion cells provided in the active region to each other. In a method for manufacturing a photoelectric conversion device, a first mask constituting a connecting portion is placed in an inactive region on a transparent substrate, and a first electrode having a transparent conductive film and a transparent conductive film are selectively formed. forming a non-single crystal semiconductor device on the active region that generates a photovoltaic force upon irradiation with light; and forming a second mask forming a connecting portion on the non-active region. selectively forming a second electrode on the non-single crystal semiconductor;
a step of connecting the electrodes of the active region to each other at a connection portion of the non-active region, and a step of separating the first electrode, the non-single crystal semiconductor, and the second electrode of the active region into a plurality of photoelectric conversion cells by laser beam irradiation. A method for manufacturing a photoelectric conversion device, comprising forming a plurality of photoelectric conversion cells in the active region and electrically connected to each other in the inactive region. 2. In claim 1, laser scribing is achieved by irradiating and scanning a laser beam from the transparent substrate side to scatter the first electrode, non-single crystal semiconductor, and second electrode in the irradiation area. A method for manufacturing a photoelectric conversion device, characterized by performing the following steps.