Method of manufacturing crystalline silicon solar cells using co diffusion of boron and phosphorus

<img file="NL2000999C2_D0001.tif" />

Patent Center

Netherlands © 2000999 © C PATENT²⁰ © Application number: 2000999 Submitted: 13.11.2007 lnt.CI .:

H01L31 / 18 (2006.01)

<td> ©</td><td>Signed up: 14.05.2009</td><td> ©</td><td>Patent holder (s): Energy Research Center of the Netherlands in Petten.</td>

<td> ©</td><td>Granted: 14.05.2009</td><td> ©</td><td>Inventor (s):</td>

<td> ©</td><td>Issued:</td><td></td><td>Valentin Dan Mihailetchi in Groningen. Yuji Komatsu in Alkmaar.</td>

<td></td><td> 01.07.2009</td><td> ©</td><td>Authorized representative: Ir. A. van Westenbrugge et al. 2502 LS The Hague.</td>

© Method for manufacturing crystalline silicon solar cells using codiffusion of boron and phosphorus.

© The present invention provides a method of fabricating a crystalline silicon solar cell, comprising in sequence:

- providing a crystalline silicon substrate comprising a first side and a second side opposite the first side;

pre-diffusing phosphorus into a first side of the substrate to form a phosphorus-diffused layer having an initial depth;

blocking the first side of the substrate;

- exposing a second side of the substrate to a drilling ditfusion source;

heating the substrate for a certain period of time and to a certain temperature in order to diffuse boron in the second side of the substrate and at the same time to further diffuse the phosphorus into the substrate.

NL C 2000999

This patent has been granted regardless of the enclosed result of the prior art research and written opinion. The patent corresponds to the documents originally filed.

The Netherlands Patent Center is an agency of the Ministry of Economic Affairs.

METHOD FOR MANUFACTURING CRYSTALLINE SILICONE SOLAR CELLS USING CO-D1FUSION OF DRILL AND PHOSPHORUS

The present invention relates to the manufacture of a solar cell using a substrate of crystalline silicon (Si). An example of such a solar cell is a borer-emitter n-base solar cell with a back side field resulting from a phosphorus diffusion.

When two kinds of diffusion processes (boron and phosphorus) are required, the higher temperature boron diffusion step is usually performed before the lower temperature phosphor diffusion step, see for example T. Buck et al., Proceedings of 21st European Photovoltaic Solar Energy Conference (September 4-8, 2006) , Dresden, Germany), pages 1264-1267. Such a process sequence requires a special protective layer to prevent phosphorus from diffusing into the boron-diffused side during the time of the phosphorus diffusion step. Sometimes boron diffuses from the boron-diffused layer to this protective layer and is exhausted near the interface. This causes an increase in the emitter's layer resistance, which leads to an increase in the series resistance of the solar cell manufactured by this process. Furthermore, additional process steps are required to remove the protective layer, or the optimum properties of the protective layer are compromised if it is retained on the silicon substrate (for example, as a passivation and anti-reflective coating).

On the other hand, if the phosphor diffusion step is processed before the drill diffusion step, the phosphor diffused side must be protected from boron during the drill diffusion step. Furthermore, sufficient phosphor must be prevented from diffusing into the drill side during the phosphorus diffusion because it cannot be easily compensated by boron. Furthermore, phosphorus escapes from the phosphorus-diffused layer at the temperature of the boron diffusion, and therefore phosphorus co-diffuses into the surface of the boron-diffused layer with boron. This hinders obtaining good properties of the boron-doped emitter. Because of these difficulties, the phosphor diffusion has hardly been attempted prior to the boron diffusion or, if attempted at all, has been unsuccessful in manufacturing a solar cell.

Although it is possible to form both boron and phosphorus diffusions simultaneously by a number of methods, such as printing each diffusion source on one side prior to the diffusion process, such a method results in compensation of boron by phosphorus at least on the edge of the drill side, because phosphorus diffuses faster and is more soluble in silicon, and therefore easily compensates for boron.

Other methods are known using separate diffusions for boron and phosphorus, with wafers placed together in pairs. Two sides of two substrates touch to partially protect them from diffusion. This will mitigate the problems of phosphor boron compensation, and vice versa, at the edge of the wafers. However, the edges of the wafers then have to be cut, which significantly increases the cost per Wp produced.

It is an object of the present invention to provide a method of fabricating a solar cell from an Si substrate using both boron and phosphorus diffusion solving at least one of the above-mentioned problems.

The object is achieved by a method of fabricating a crystalline silicon solar cell, which method successively comprises:

- providing a crystalline silicon substrate comprising a first side and a second side opposite the first side;

pre-diffusing phosphorus into the first side of the substrate to form a phosphorus-diffused layer having an initial depth;

blocking the first side of the substrate;

exposing the second side of the substrate to a drilling diffusion source;

heating the substrate for a certain period of time and to a certain temperature in order to diffuse boron in the second side of the substrate and at the same time to further diffuse the phosphorus into the substrate.

The present invention involves stabilizing and reducing the separated phosphorus in the atmosphere during the boron diffusion process, by already diffusing the phosphor into the surface to some extent before the boron diffusion. This makes it possible to remove the phosphorus diffusion source before the drilling diffusion. The amount of phosphorus separated from the diffusion source is greater than and fluctuates more than that of the silicon surface where phosphorus has already diffused.

Therefore, this process improves the quality and reproducibility of the boron-diffused p-type emitter. It also avoids the formation of an n-type edge on the boron-diffused side and therefore prevents shunting of the solar cell.

In one aspect, the invention also relates to a solar cell manufactured by the method described above.

Further advantages and features of the present invention will become apparent from a description of a number of embodiments referring to the accompanying drawings, in which:

Figure 1 shows an example of the structure of a solar cell produced by a method according to an embodiment of the invention;

Figure 2 shows another example with a p-type silicon substrate;

Figures 3A-3D schematically show processing steps for making a P-diffused layer on one side of a substrate;

Figures 4A-4F schematically show processing steps of an alternative method of fabricating a P-diffused layer on one side of a substrate;

Figures 5A-5D schematically show processing steps for a possible third method of fabricating the P-diffused layer on one side of the substrate;

Figures 6A, 6B show two possible configurations for blocking one side of the substrate during a drilling diffusion step;

Figure 7 schematically shows a back-to-back configuration during a drilling diffusion step;

Figure 8 is a graph showing the measured efficiency of solar cells produced by the method of the invention compared to prior art solar cells.

Figure 1 shows an example of the structure of a solar cell produced by a method according to an embodiment of the invention. A solar cell 10 includes an n-type silicon substrate 11 that includes a boron-diffused layer 12 on one side and a phosphor-diffused layer 13 on the other side. It is noted that a practical solar cell structure also includes metal contacts and an anti-reflective coating, but these components are not shown in the figures. Figure 2 shows another example in which a p-type silicon substrate 21 is processed to produce a phosphor-diffused layer 22 on one side and a boron-diffused layer 23 on the other side. The solar cell of Figure 1 is the preferred embodiment because its device performance is better than that of Figure 2. In the description below, embodiments of the manufacturing method of the solar cell shown in Figure 1 (i.e. n-type substrate) are discussed.

The first step of this method is to make a P-diffused layer on one side of the substrate. In one embodiment, a substrate 30 is heated to 800-900 ° C for 5-50 minutes in an atmosphere comprising an O2 and P2O5 vapor. Then all surfaces of the substrate 30 are covered with a S1O2 film 31 comprising P2O5 (hereinafter, SiCh2 Os). This S1O2 grows from Si of the substrate 30 and oxygen, and P2O5 is included in the S1O2 film 31. At the interface of silicon 33 and the SiCT® Os film 31, P2O5 is reduced in P and P diffuses into the core of the substrate 30 (see core 33 in Figure 3B) to a depth of 0.01-1.0 µm. So far, the SiCh 2 Os film 31 and a P-diffused layer 32 have been formed on the entire surface of the silicon substrate 30. Then, the SiCh 2 Os film 31 is removed by immersing the substrate 30 in a 1-50% HF solution for about 0.5-10 minutes, or by exposing the substrate 30 to an HF vapor, or by etching using reactive ion etching, see Figure 3C. Then, the P-diffused layer 32 is etched, except for one side, using a mixed solution of 1-30% HF and 0.1-50% HNO3, or by etching it using reactive ion etching. Etching on one side is possible by sealing the other side of the substrate 30 with an etch block coating or by just floating the substrate 30 on the solution. As a result, the substrate 30 now includes a P-diffused layer 32 'on one side, see Figure 3D.

An alternative method for fabricating a P-diffused layer on one side of a substrate is explained with reference to Figures 4A-4F. First, one side of the surface of a substrate 40 is coated with any of liquid, paste or gel 41 comprising P2O5 and S1O2 fine particles by either spin coating, spray coating or printing, see Figure 4B. The other sides of the substrate might as well be coated, but this does not affect the end result of this manufacturing method. Then the coating 41 is heated to 250-500 ° C. The solvent evaporates or burns out when it contains organic matter. The P₂0s and SiO? remain in the coating, see Figure 4C, which shows a coating 41 '. Then, in a further heating step, the substrate 40 is heated to 800-900 ° C for 2-50 minutes. As a result, all surfaces are covered with S1O2.P2O5, see Figure 4D. Atmospheric oxygen, and P2O5 from the first coating film 41 'emitted into the atmosphere. The P diffuses into the Si core just as in Figure 3B, and a SiCh 2 Os film 42 and a P diffused layer 43 are formed on all surfaces of silicon substrate 40. Now, the SiC 2 Os film 42 is removed using a 1-50% HF solution or other known method. The P-diffused layer is etched, except for the side that was first coated, using a mixed solution of 1-30% HF and 0.1-50% HNO3, or reactive ion etching. Etching on one side is possible by sealing the other side with etch blocking coating or by just floating the substrate on the solution. The result is shown in Figure 4F, which depicts the substrate 40 comprising a P-diffused layer 44 on one side.

A possible third method of fabricating the P-diffused layer on one side of the substrate is explained with reference to Figures 5A5D. First, one side of a substrate 50 is blocked using a diffusion blocking layer on one side, see blocking layer 51. The blocking layer 51 can be formed using various processes listed below:

• Coat the surface with a liquid, paste or gel that includes S1O2 or T1O2 or anything that does not diffuse into silicon by spin coating or spray coating or printing. Heat the coating to 200-700 ° C, then the solvent evaporates.

• Heat the substrate 50 to 850 - 1100 ° C in an O2 or O2 + H2O atmosphere for 0.5 - several hours. Then a SiO2 film thicker than 0.1 µm is formed on all surfaces. Remove the film on one side only by immersing the substrate 50 in a 1 -10% HF solution.

• Apply&gt; 0.1 µm thick S1O2 or SiN or T1O2 or the like using chemical vapor deposition.

In a next step, P is diffused into a Si-core 54 using the method as described with reference to Figures 3B or 4D. A

SiO2: P2O5 layer 52 is formed and P diffuses into it, but the blocking layer 51 prevents the P from diffusing into Si core 54 on one side, see Figure 5C. Then, the SiCl 2 Os layer 52 and the blocking layer 51 are removed by immersing the substrate 50 in a 1% ~ 50% HF solution.

According to another embodiment, the diffusion of P on only one side of the substrate is achieved using a back-to-back diffusion method in which two substrates contact each other at their surface.

After the pre-diffusion of phosphorus into the first side of the substrate described above, that same first side of the substrate is blocked before the substrate is placed in an oven for further processing. In one embodiment, the first side 61 of the substrate 60 is blocked by a first side 62 of another substrate 63. That other substrate may be a similarly treated substrate, see Figure 6A. This blocking method is referred to as back-to-back. One of the advantages of a back-to-back configuration is that less space is required in the oven compared to individually blocking each substrate. Furthermore, the escape of phosphorus from the first side 61 is prevented very effectively, because the substrate turned to it also has a high phosphorus concentration, so that the phosphorus concentration is kept in better condition. Figure 6B shows an alternative in which the substrate 60 is blocked by a substrate 65 that has not yet been treated (i.e. a new substrate).

In the oven, the second side of the substrate is exposed to an on-board diffusion source. This boron diffusion source can be a vapor source or a coating source. In the oven, the substrate is heated for a certain period of time and to a certain temperature, in order to diffuse boron in the second side of the substrate and at the same time to diffuse the phosphor further into said substrate (i.e. deeper than said initial depth ). Successful results have been achieved with a boron vapor source for the diffusion. Below, a specific description of an embodiment is described with reference to Figure 7. Two substrates 70, 71 are placed back to back in the oven and heated to 900-1000 ° C for 30-120 minutes in an atmosphere containing an O2 and B2C &gt; 3 vapor, which can be produced by passing N2 through BBr3 liquid. Other boron fluid sources such as BCI3 or tremethylborate can also be used in place of BBr3. Then the exposed surfaces (i.e. the surfaces that are not blocked) are covered with an S1O2 film 72 comprising B2O3 (hereinafter, SiCb ^ Cb). At the interface of Si cores 70, 71 and the SiCbiELCh film 72, B diffuses in Si to the depth of 0.01-1.0 µm to obtain B-diffused layers 73, 74. Some of the B2O3 may slip into the narrow gap between the substrates 70, 71, but the influence is very small due to the existence of heavily diffused P to those regions. At the same time, the P present in the P-diffused layers 76, 77 also diffuses further into the respective Si-nuclei 70, 71, driven by the heat used. This will lead to P-diffused layers deeper than their original depth.

The drill diffusion from the example of Figure 7 utilized vapor source diffusion, it will be apparent to those skilled in the art that this method is also effective in the case of a coating source diffusion as described in the phosphor diffusion step of Figures 4A-4F.

The combination of a pre-diffusion of phosphorus with a further diffusion during the simultaneous diffusion of boron and phosphorus results in a solar cell that has very good properties, as can be seen from figure 8. Figure 8 shows a graph of efficiency values ​​of solar cells are fabricated on n-type multicrystalline substrates using the novel method (as presented here) and the prior art method (see, e.g., T. Buck et al., Proceedings of 21st European Photovoltaic Solar Energy Conference (September 4-8, 2006, Dresden, Germany), pages 1264-1267).

As a consequence of the manufacturing method described above, phosphorus is diffused more deeply than when phosphorus is diffused individually (not in a simultaneous diffusion with boron). When the invention is used, a phosphorus concentration in the substrate at 0.5 μτη depth can be more than 100 times higher than at 5 µm depth.

In the method of the invention, phosphorus can also diffuse into the silicon on the drill side because a small amount of phosphorus will flow from the phosphor side to the opposite side (i.e., the boron-diffused side). On the drill side, however, the amount of diffused phosphorus is less than the amount of diffused boron, and the diffused layer can easily meet the proper conditions for a p-type emitter.

The amount of phosphorous phosphorus is still greater than the background doping of the substrate. The phosphorus concentration at 0.2 µm deep can be more than 100 times greater than at 5 µm deep.

The invention allows the manufacture of a boron dopant profile that meets the required conditions for an emitter, without allowing phosphorus of a greater amount than boron to diffuse into the boron-diffused side and without allowing boron of an amount greater than phosphorus diffuses into the phosphorus-diffused side.

It will be apparent to those skilled in the art from reading the above text that variants will occur. These variants are intended to be within the scope of the invention as set forth in the appended claims.