JPS61500757A - How to manufacture solar cells - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] This application is a successor to U.S. Patent No. 563,132, filed December 19, 1983. No. 666,973 filed October 31, 1984 This is a continuation application.

This invention relates to the production of photovoltaic cells, and more particularly to the production of photovoltaic cells, and more particularly to Polycrystalline silicon layer used as plating mask for surface electrode metallization It concerns an improved and inexpensive method of manufacturing solar cells.

Conventional methods for manufacturing silicon solar cells typically involve the production of silicon wafers or ribbons. The step of forming a PN junction is done by diffusing an appropriate amount of t1-pant into the surface. A grid pattern is engraved on the protective insulation coating 1171 formed on the surface of the floor. Nickel plating on all silicon exposed by etching. The step of dip-soldering or plating copper and soot onto this nickel surface. 7) Removing the remaining portion of the insulating liquid wJ layer from the surface as well as the newly exposed portion of the surface. The method includes a step of applying an anti-reflection coating to the portion.

Although such plain silicon can be applied to either single crystal or polycrystalline silicon, Considering cost, it is desirable to manufacture solar cells from polycrystalline silicon.

However, as is well known, minority carriers are lost at grain boundaries, dislocations, etc. For this reason, the efficiency achieved with polycrystalline silicon solar cells is comparable to that of monocrystalline cells. It's not good for This situation can be solved by introducing -valent elements, such as hydrogen, into the crystal structure. Combining with defect-related dangling bonds to minimize recombination loss of minority carriers. It has been improved by minimizing the

As is known in the art, important considerations when designing a photovoltaic fabrication process are: The time and temperature combinations in any step following the hydrogen passivation step in silicon This is because the hydrogen introduced into the substrate does not diffuse backwards from the inactivated substrate. Ru. That is, for example, hydrogen dehydrogenation exposed to a temperature of 600°C for half an hour in vacuum. Activated photovoltaic cells are evidenced by their observed electron beam-induced current activity rates It has been found that almost all bonded hydrogens are lost as shown in FIG. this It should be noted that the metal diffusion step in solar cell manufacturing is typically Typical: It requires a temperature of about 900°C.

Hydrogen passivation is usually done by heating the photovoltaic cell to a sufficiently high temperature to form a base metal Rk junction such as copper. It is also possible to diffuse into the It's clear. For example, "Journal of Vacuum Science and Technology" cuum 5 cienreand Technology)j 2nd f) Winding cylinder In No. 3 430-435 <-ji (19”! March issue 2) Gar (C, H.

Seagsr), D.G. Sharp (D, J, 5harp) *Jie I.K.G. Panic (J, Ni, G, Pan1tz) and R.B.I.D. Ix　o　/R,V,D'A15Llo) Passivation of silicon generates hydrogen ions in the kiloelectronvolt energy range This can be done using a Kaufmann type ion source used in High ion energy energy and flux (e.g. 1 to 3 milliamps per square centameter) A relatively short exposure time in the area (eg 0.5 to 4 minutes) appears to be optimal. Ru. Such exposure generally occurs when the board is carefully brought into contact with a suitable heat sink. In this case, the substrate temperature will be increased to at least about 275°C. That's right If not, temperatures of 400° C. are easily achieved. However, Siri The temperature is controlled to below approximately 300℃ to avoid rapid diffusion of the frame metal into the concrete matrix. It is necessary to limit N. However, the substrate for thermal control during passivation and the heat dissipation element is easy to operate (C. This will cause a decrease in treasure. Therefore, in order to obtain a low-cost, high-speed processing process, it is recommended to use a heat sink. It is advisable to avoid it. Furthermore, EFG type/recon which can be produced economically In the case of ribbon, the uneven surface makes it difficult to use the heat sink.

Furthermore, hydrogen passivation is most effective when the underlying silicon surface is exposed. Ru. Therefore, any mass, such as the silicon nitride layer used in conventional methods, must be removed prior to inactivation. Therefore, the surface electrode is If it is not applied before, it will not be '5i.

Described in U.S.A. 2/Application No. 563061 (Agent's Document No. MTA-49) As shown in the figure, the altered surface layer generated during hydrogen ion beam passivation is Can be used as a plating mask for metallization steps. In more detail , as described in detail in the above-mentioned application as applied to the production of silicon solar cells. Selected embodiments of the method include, inter alia, the following steps. In other words, (1) Shallow A dielectric material plating mask is formed on the surface of the silicon ribbon for bonding, and then a surface electrode is applied. Be sure to leave exposed the part of the silicon that will be covered by the pole. (2) depositing a thin layer of nickel (or similar material) on the exposed silicon; (3) removing the plating mask, (4) joining the ribbon Step of hydrogen inactivation on the side, (5) Nickel? Sintered and partially silicified chive (6) plating additional metal on the metallized portion of the photovoltaic cell; and (a step of applying an anti-reflection film to the exposed surface of the silicon. After that, the silicon The condenser may be further processed to enable it to be connected to an electrical circuit, for example. cormorant. Passivation alters the exposed surface on the bonding side of the substrate and this is the secondary plating step (6) To serve as a mask for.

In another method, heating the material during passivation is an energy source for the nickel sintering step. supply at least a small portion of Rugy.

This thousand roses are U.S. Patent Application No. 563292 (Agent's document number MTA-50) In this application, the following is applied to the surface of the substrate. Negative plating mask (i.e. the surface that will later be covered by surface electroplating) A mask that covers the surface area (Mi1) covers the altered surface layer generated by the ion beam. This can be used to control the range, and this can be used before any surface metallization. If passivation is possible, the mask is removed after passivation and before metallization. are taught and the 1,000-year method just outlined both enable hydrogen inactivation. , which results in improved photovoltaic performance of polycrystalline substrates under desirable temperature conditions. However However, even the simplest of these two foolproof steps outlined takes about 12 minutes. It requires several steps. Both procedures additionally provide a mask in which the surface layer of the substrate can be temporarily disposed of. This requires time and non-reusable materials to be formed. Eliminates processing steps when manufacturing solar cells from substrates similar to EFG type Lipolybon. It is the purpose of this invention to do so.

After the high temperature treatment step but before any base metals are incorporated into the crystal structure, hydrogen depletion occurs. Process 1 for manufacturing a solar cell included in activation stage '4, and Another aspect of the invention is to provide one that minimizes the use of non-recyclable materials. It is a purpose.

Detailed description of the invention These and other purposes include the application of ions on the surface of the substrate to be used for hydrogen deactivation. A removable and reusable mechanical mask is used to cast the beam shadow. This invention is satisfying. The ion beam passing through the mask aperture The altered surface layer generated during this process forms the boundary between the areas of subsequent surface metallization by displacement plating. Form the specified plating mask.

As may be noticed, delimiting the portion of the substrate exposed to the ion beam Use of a removable mechanical mask to remove A stubborn plating mask? Chemical milling steps used to form and remove Second, the mask of this invention is reusable. The result is thicker. A simplified and more cost effective process is enabled.

The features of this invention, E:U, will be partly obvious, and partly will be explained below. It will become clear. This invention therefore has the following detailed disclosure: Some of the steps illustrated and one or more of these steps and others The scope of this invention is shown in box Q of the claims. For a more complete understanding of the nature and purpose of the invention, reference may be made to the following detailed description. However, this description does not cover the manufacturing of solar cells according to the adopted form of this invention. Many stages with weak threads? Should be considered in conjunction with the accompanying drawings shown. be.

Throughout the drawings, the aircraft reference numbers indicate similar structures.

In the drawings, the thickness and depth of some layers and regions are shown for illustrative purposes only. are not precisely illustrated according to their relative proportions.

Detailed description of the invention Now to talk about the other side, the adopted embodiment of this invention is an EFGa long P-type laminar It is concerned with the production of large-scale ponds from ribbons.

Initially, as a process requirement, 6 EF (, P type conductive 7 recon ribbon) is pre-cleaned and heated. One side of 2 (hereinafter referred to as the "front side") has a relatively shallow junction 4 (that is, approximately 3Q). an N-type conductive region 6 A carefully calculated diffusion process to produce receive. -For example, by EFG method A P-type conductive silicon ribbon is manufactured and has approximately 50 near 1 resistors. is -'J4:1 to 9:1 ratio (7) HNO3 (70%): HF ( Etching was performed at approximately 25°C for approximately 3 minutes in a molten a of 49%). Therefore, it is purified. The ribbon is then exposed to an oxygen-rich atmosphere, as is well known in the art. Receive phosphorus diffusion treatment in air. For example, as detailed in U.S. Pat. No. 4,152,824, Phosphosilicate glass 8? Rind -Can be used as a source for mint.

Next, a layer of phosphosilicate glass 8 is applied by dipping the base into a buffered solution (and thereby engraving it). be removed. For example, (P2O3)X(Si02)Y,') phosphorsilicate glass (・Substrate klONH4F (40%): In a solution of IHF from about 25°C to about 40% Removal from the substrate by incubation for approximately 15 to 2 minutes at a temperature of 0°C. Can be done.

Following this, the back side of the substrate is coated with an aluminum paste, 110. layer The aluminum paste used to form 10 is removed by evaporation. Volatile compounds such as terpineol obtained from aluminum powder in the color vehicle It is desirable that

This step is followed by a further alloying step in which the substrate is approximately zero. ．． 25 rL about 575°C for 2.0 minutes: high temperature I' this 11, ton roundness All volatile or heat-labile components of the paste are removed and the paste is removed. The aluminum in the process is alloyed onto the silicon substrate.

In the alloying stage, the aluminum coating 71O forms an alloy with the back side of the substrate, forming an alloy of about 1 An fcP+ region 12 having a depth t of 5 to 5 microns is shaped and cut out.

Next, the -yt battery (cell) is hydrogen activated. The adopted method was Hydrogen ion beam of Kaufmann type (wide beam) ion source located at jJ 15cmL. The surface of the substrate 2 is exposed to the beam. A mask 14 is provided between the ion source and the substrate. It has been done. This ion source should preferably be about 20 Pressure of 50 mTorr, about 25 to 40 s, c, c, hydrogen per minute Flow bed, potential of about 1700 volts DC, and about 1 to 3 milliamps/cIn2 operated with a beam current of Exposure (irradiation) time of about 1 to about 4 minutes is EFG type. Is minority carrier recombination loss a sudden surprise in photovoltaic cells? Minimally about 20 The passivation region is 80 microns deep, or approximately 100 times deeper than junction 4. ) and at the same time a modified surface layer 1 of approximately 200 angstroms on the exposed portion of the substrate 2. 8? It turns out that there is enough to give.

The positive quality of the altered surface layer 18 is unknown. However, that, well, crystal The structure is somewhat destroyed and the 7 recon is partially exposed to hydrogen from the ion beam and SiH or 5IH2 shape, but in damaged areas where the material is possibly amorphous. It is believed that there is. A small amount of carbon or one or more hydrocarbons forms the desired altered surface layer. As I grow older, I forget that I need it. Where it was first installed, it was used The Kauffman ion source is equipped with a graphite mount approximately 5 inches in diameter. with a typically 2 x 4 inch (5 x 10α) surface in the center above it. The board is placed. In some cases, a 7-recon mount can replace a graphite pedestal. When a graphite table is used, the altered layer is It did not function well as a school.

A hypothesis that has been touted based on this/ζ is that the hydrogen ion beam collides with the graphite table. The scorched hydrocarbon vapor formed into a shape 1 forms a dielectric material 1 on the surface of the substrate. I think you can. No matter what Cao is, it is about 140 years old according to this plain. Made with an accelerating voltage of 0 to about 1700 volts and an exposure time of 1 minute. The altered surface layer 18 prevents subsequent damage to the substrate on this layer. is sufficient.

Mask 14i has the desired multi-fingered grating structure, as described in U.S. Pat. No. 3,686,036. An example of this is the gold '14'A reticle in the fc electronic pattern. . Mask 14'I'i is preferably made of molybdenum, but other metals such as Imphal, stainless steel, titanium, nickel, etc., or graphite and similar high temperature non-alloys. It would also be possible to use metal materials. The mask 14 covers the surface of the substrate 2 and ions. For the ion beam source, when the ion beam source is activated Shadow of the desired grid pattern? It is distributed as if it were dropped. In other words, 'i, mass 141. Once the desired MIG grating pattern is set, the beam 16 is completely blocked 9 and the electric power ( In 9 parts, it is possible to irradiate the board. The metallization of the photovoltaic cell then takes place. Both sides of the board are displacement plated with nickel. Kell's ear settlement N Nickel layer 22? The shape and the protective adhesion of nickel on the front side are directly applied to the substrate 2. ．． 7) The shape of the nickel layer 20 is changed only in the area where there is no altered layer 18 on the surface. child In the plating stage, the 7 ReconD altered surface 118 is not marked with nickel. Tsufi, VL mask? Form. The plating of the nickel layer can be done in various ways. I can do it. Preferably, it would be Kirit Patel. sl) or the method described in US Patent No. It is carried out according to a displacement plating method similar to .

As used here, the term "one-replacement" refers to a method that does not contain a reducing agent. By immersing the object in a plating bath, it is possible to target this object without using an externally applied electric field. What do you do? Therefore, this plating necessarily involves a conversion reaction. With this:i In contrast, the term "electroless plating", well, reducing agent? In the plating bath containing: Plating performed without an externally applied electric field by dipping the object to be plated? means Ru.

As a reserve, the cleaned 7 recon board surface is pre-activated with a suitable chemical. Ru. This preactivation process often makes the silicon surface itself resistant to electroless plating. nickel plated on untreated surfaces generally adheres poorly. So, desirable. Preferably, gold chloride is used as the activator, but platinum chloride , palladium monochloride, or the well-known activators of It is also possible to use one such as that described in Japanese Patent No. 3,489,603. So After that, both sides of the silicone ribbon are preferably It is coated with a layer of nickel by soaking for 6 minutes.

After being nickel-plated, the substrate is exposed to an inert atmosphere or an atmosphere such as nitrogen and hydrogen. In the mixture, sinter the nickel layer and nickel layer 20 on the front side of the substrate? adjacent The temperature is high enough to react with silicon to form ohmic contacts of nickel silicide. It is heated for a long time. For this purpose, the substrate preferably has a fc of about 300 ℃ for about 15 to about 40 minutes. Nickel like this: vI2 nickel silicide to a depth of approximately 300 ogstroms at the interface between substrate 2 and substrate 2. ・1 is formed. , the nickel layer 22 on the back side forms an alloy with the aluminum layer 10 do. This rotation of Samurai Yakugaya should not exceed 300 degrees Celsius. SoiL Yoji High temperatures also result in a poor quality nickel layer 20i, as previously mentioned. Is it possible for some of the passivating hydrogen to diffuse back from the substrate material? It is et al. Preferably, the nickel deposition and sintering should be performed on the nickel on the front side of the substrate. The QiJ value is such that the layer has a thickness of less than about 750 angstroms.

The nickel in layers 20 and 22 is then etched, such as by etching. Additionally, a second layer of nickel and one or more layers of copper may be added. It is desirable to receive it. Additional nickel (preferably by displacement plating) , is molded in the manner previously described for the formation of layers 20 and 22; L: Nickel is plated on 20 and 22 at the mountain part of the altered layer 18. This is because Preferably, the sea bream should be caught using the well-known technique: 9. and/or by electroplating.

After the gold conversion, the q section (not shown) of the photovoltaic cell is trimmed and the anti-reflection coating 24 is applied. is applied to the surface of the photovoltaic cell. This can be done by many known methods For example, chemical vapor deposition (CVD) or vapor deposition of TiO2 may be used.

Or, as is well known in the art: plasma of silicon at a temperature of about 150°C. The antireflection pattern 24 may also be formed by adhering the material.

As an example, the adopted method of carrying out this invention: well, the above-mentioned lock step , in accordance with the order outlined in the adoption form, which explains each step in detail. It consists of things.

EFG grown ribbons are made from Hokuten and Natsu 7 ponds according to the method described above.

Furthermore, for this material, the hydrogen passivation step also significantly changes the distribution range of photovoltaic efficiency. Made it narrower.

It will be appreciated that the method of this invention greatly facilitates the production of high-temperature batteries. . This method not only reduces the number of process steps required, but also reduces the amount of recovered waste. Eliminate the use of possible materials. In other words, photoresist is used to cover the top of the surface layer. Is it possible to make a plating mask in the shape of 5y? By eliminating the Photoresist like? The need to subvert, expose and develop? At the same time as eliminating Rino photoresist? Removing it eliminates a necessary step. Similar worries (hei is chemistry) Equivalent process of Cal/Ring (But, Atari, Mamaru.

Can the method described above be modified without departing from the Q of this invention? It will be understood. In other words, the adoption of the method of this invention i, rl, l: Alteration/Jin formed by hydrogen inactivation? Previously I used Notsukisa Kuta Nikke. This method uses metals other than nickel as a mask for subsequent plating. It can also be used for For example, in order to be understood by those who are technically savvy, Initial layers such as the surface layer in a shallow junction 7 recon element (preferably at a low temperature) copper or other base metal that forms an ohmic contact and is deposited at a later stage Any of a number of low reactivity materials that can serve as a barrier to the diffusion of It may be attached by replacing W. gold suitable for use with copper Beak is nickel: From the source, ξ radium, platinum, cobalt 1. and rhodium Contains. These materials are combined to form silicides, but silicides and blinds are necessary. Not essential. However, the initial gold 4 layer (adhered properly and returned ohm) It serves as a barrier to diffusion of any metal that is later deposited. It is important that the bonding agent not only act as a protective layer but also not diffuse too much into the bonding itself.

Other changes may be made without departing from the present invention's cylindrical blade), such as (α) aluminum p+-JE side of photovoltaic cell using flame exposed aluminum instead of aluminum paste or (b) nickel or 11, palladium, platinum, copper. 7) Low reactivity materials such as rhodium 87.1 Second and subsequent coatings? Execution It is also possible to form the junction by using another method called (・ma(C)) ion implantation method. It can be done by shouting. Paraphrase 1 above (so that the mask layer is not attached) Additional nickel (or other low reactivity in the form of a metal) shall be applied by one-pass plating.

Of course, the method provided by this invention allows for the production of solar cells 7) from EFG substrates. It is not limited to construction. For example, cast polycrystalline substrates, metal handling/recycling, etc. Epitaxial 7-year recon, or chemical or chemical vapor deposition on the condenser: ζyotsu In this invention, the C-formed fC high-purity poly/recon 7 layers (Niji relatively high efficiency solar Can be used to charge all types of batteries. Furthermore, this method can be applied to single-crystal silicon. is also applicable. And, this method also applies to P-type silicon, as well as N-type silicon. can also be implemented.

In both of these cases, bonding is not simply due to phosphorus diffusion; It will be appreciated that it can be formed in the following manner.

These and other methods may be incorporated into the foregoing without departing from the scope of the invention disclosed herein. Other changes may be made and are not included in the foregoing description or in the accompanying drawings (see C). All matters herein shall be construed in an illustrative rather than a restrictive sense. This is considered to be the case.

International tIl Sai Report

Claims (9)

[Claims] 1. (a) providing a silicon substrate with opposing first and second sides; (b) placing a mechanical mask having selective apertures in juxtaposition to said first surface; every day (c) said first surface corresponding to said selected opening and to which the metal plate is only incompletely attached; Water through said selected apertures at sufficient strength and time to form a top surface layer. exposing the first surface to an elementary ion beam; A method of manufacturing a solid-state semiconductor device including in order. 2. Further, after the exposure to the hydrogen ion beam, the mask is covered with the mask. 2. The method of claim 1, wherein portions of the surface that have been coated are metallized. 3. According to claim 2, all of said surfaces are exposed to metal during metallization. How to put it on. 4. a front surface adjacent to said first surface before exposing said first surface to said hydrogen beam; The method of claim 1 further comprising forming a bond to the substrate of claim 1. Law. 5. sufficient time and intensity to reduce minority carrier loss in said substrate. Claim 1, wherein the first surface is exposed to the hydrogen ion beam at The method described in section. 6. The metallization includes nickel, palladium, cobalt, platinum, and rhodium. The method according to claim 2, which is carried out using a metal selected from the group of metals including Law. 7. placed for metallizing the part of said surface covered by said mask. 2. The method of claim 1, further comprising the step of performing a replacement. 8. said metallizing step is performed to also metallize said second side of said substrate. 8. The method according to claim 7. 9. The above metallization involves the formation of nickel from a solution containing nickel salts and fluoride ions. 3. A method according to claim 2, comprising deposition.