FR2472840A1 - Solar energy cell made from silicon wafer - where surface of wafer is etched to produce recesses reducing reflection and increasing light absorption - Google Patents

Abstract

One surface of the wafer is employed to receive and absorb incident light and to convert the light into electrical energy. The surface is first coated with a mask which resists the attack of an etchant, and the exposed zones of the Si are then etched to produce symmetrical recesses in the surface, which is subsequently provided with an electrical connection. The mask pref. has a uniform thickness and covers all the surface of the Si, the mask is made esp. by an agent which oxidises the Si to form silica, and the silica is pref. covered by photoresist and etched to form the mask. The recesses are pref. inverted pyramids, of which the bases occupy more than 90% of the surface of the Si. The rear side of the wafer may also be provided with similar recesses filled with metal acting as a heat sink. All the incident light is absorbed, none being reflected, so the energy yield of the cell is increased.

Description

 The present invention relates generally to solar-powered silicon cells having at least one main surface for receiving and absorbing the light incident on it. More particularly, it relates to cells used to produce electricity in terrestrial atmosphere although it is also applicable to cells used as energy source for orbital satellites and other devices located in space and which therefore does not not benefit from an atmosphere to attenuate the bombardment of the cell by particles of energy.

 There is a problem known for many years to specialists in this type of solar cell and which has not been satisfactorily resolved, this problem being to minimize the reflection of light striking a light absorbing surface of a solar cell and increase absorption. Although anti-reflective coatings composed for example of tantalum pentoxide or niobium have proved useful in this respect, it has been found that cells provided with such coatings absorb useful light and do not offer complete effectiveness for all lengths useful waves.

 Another way to increase the absorption of available light, and which has been the subject of some interest, has been to texturize the light impact surface of the cell. In a particular application of this technology the cell surface is attacked by a solution of potassium hydroxide or sodium hydroxide to create a cell surface in which pyramids projecting upward are formed. These pyramids are of random sizes. The purpose of these pyramids protruding upwards on the surface of the cell is to capture the light, i.e. the light which is not absorbed by the cell on a main surface is fortunately reflected on another structure pyramid or on the main surface and is absorbed by the surface or the reflection pyramid.

 Trapping light by means of upwardly projecting structures on the surface of a solar cell has certain disadvantages, although these structures in principle increase the absorption of available light. Among these disadvantages is the fact that the pyramid tips or parts of any structures projecting upwards on the surface of the cell can be broken during handling. Since a cell having a textured surface is subjected to impregnation using a junction-forming impurity after the structure has been made, if the tip of a pyramid subsequently breaks, the part of the cell where the tip is originally located will not have a photovoltaic junction. As a result, the efficiency of the cell will be reduced, and an electrode metallization design in the form of a grid is difficult to apply due to protruding algae tips. Thus the metallic grid pattern in certain cases must connect the neighboring points which makes the grid fragile. In addition to the fact that at least the base metal of the metallized pattern is usually applied by "shadow mask" masking methods, the application of the mask to the surface of the cell can break the tops of the pyramids projecting towards the top, in which case double masking is necessary, otherwise the grid pattern will be discontinuous. The main disadvantage may be that when the metallic grid pattern contacts a pyramid whose apex has been broken, which exposes regions p and n, the grid short-circuits the cell. These difficulties, in particular the probability short-circuiting the cell has had the effect of hampering the wide development of structured surfaces in the field of photovoltaic cells.

 From another point of view, solar cells have constituted an important part of the space programs of the United States because of their use as the main source of electrical energy for space vehicles. Thus programs have been developed to increase the efficiency of cells and have thus led to reduced costs and improved operation of space missions where vehicles are powered by solar cells. However, the most recent work in the development of silicon solar cells has focused on improvements in efficiency at the start of cell life. This is certainly an important factor and leads to cells which are more economical for terrestrial use. However in space the cells are bombarded by particles of energy not attenuated by the atmosphere surrounding the Earth. . As a result, the cells that are used for the space program have been found to have a significantly shorter lifespan than those used only for land uses. However, it is well recognized that the solar cells used to power orbital satellites have a useful life of around 7-10 years. This is why the useful life of a satellite has normally been limited to that of the cells which supply it with energy, that is to say approximately 7 to 10 years. For some time continuous efforts have been made to improve the radiation resistance of solar cells.

Among the efforts that have been made to improve the resistance to radiation of solar cells by changing the physical form of the cell junction is what has been called the vertical junction cell. Such a cell was the subject of a scientific article entitled "New Development in
Vertical-Junction Solar Tells, presented at the Twelfth IEEE Photovoltaic Specialists Conference held from 15 to 18
November 1976 in Baton Rouge, Louisiana. In general the purpose of making vertical channels in the 1-0-0 surface of the solar cell which results in the photovoltaic junction extending vertical channels inward from the cell surface is to bring the junction more close to the light energy absorbed by the cell. Because defects in the cell structure caused by exposure to the outside space can reduce the efficiency of light-carrying carriers, reducing the distance between the carriers and the junction increases the probability that these carriers reach the junction.

 However, cells with a vertical junction while being incontestably advantageous as regards resistance to radiation, that is to say an ability to function in spite of damage, have the disadvantage of being fairly difficult to manufacture in a precise manner. In the current state of knowledge, they can only be produced in a surface on the 1-1-0 plane of monocrystalline silicon in which the common planes are 1-1-1 and 1-0-0. As a result, vertically junction cells are difficult to reproduce uniformly and are relatively expensive to manufacture. However, there is no doubt that grooved solar cells with vertical junctions have a degradation rate by irradiation significantly lower than those of silicon solar cells in which the junction is planar. With regard to the increased absorption of light, this increase can be obtained when the walls separating the grooves are very thin, but then the structure is extremely fragile.

 The solar cell which is the subject of the present invention uses some of the advantages of the vertical junction cell, with its improved resistance to radiation damage, and some of the textured cell with its increased absorption but without its tendency to be easily subject to physical damage. The main object of the present invention is thus to produce a solar energy cell having a surface suitable for receiving and completely absorbing the light incident on it, said cell, better resistant to damage by radiation than a cell having a planar junction.

 More precisely, a solar energy cell according to the present invention has a main surface adapted to receive the light incident on it, said surface being provided with a plurality of indentations whose bases lie in the plane of the main surface. These indentations generally define voids having the shape of a geometric figure, in fact a pyramid.

Generally the largest area of the figure in any section parallel to the main surface will be on the surface of the cell. Consequently when the geometrical figure is a pyramid the body of the cell delimits voids having the shape of pyramids whose points extend vertically inwards from the surface of the cell and whose bases are in the plane of this area. Preferably the pyramids are distributed uniformly over the surface of the cell and occupy at least 90% of this surface.

 As a result of the use of a cell having the structure which has just been described, that is to say having a series of voids in the form of inverted pyramids formed in the surface of the cell, a cell is obtained which resists radiation and which, when it reflects light, probably reflects it on another of the surfaces formed by the pyramid so that the light energy can then be absorbed by this surface. In addition, the disadvantages inherent in textured cells with upwardly projecting pyramids are avoided; there are no upward protruding structures that could be broken or otherwise altered during manipulation of the cell.

Thus the ability of the cell surface to reflect light to another absorbent surface is maintained in the manner of a cell with pyramids projecting upwards.

 An important advantage of a cell according to the present invention is that it lends itself to a relatively simple production process while being efficient. This process uses the ability of certain etchants to easily attack silicon while not satisfactorily attacking certain other materials. For example, it is well known that potassium hydroxide and sodium hydroxide attack silicon but do not easily attack other materials such as silicon oxides. Thus the method according to the invention consists in masking the surface of a silicon wafer with a layer resistant to attack by such a silicon attack agent, in forming open regions in the masking layer in order to expose portions substantially symmetrical to the surface of the silicon wafer below these regions. The masked surface is usually the 1-0-0 silicon surface. Then a silicon attack agent is applied to the masking layer including its open regions. The silicon attack agent penetrates through the open regions into the masking layer and makes contact and attacks the surface of the wafer, forming pyramidal structures or other geometric structures directed towards the inside of the surface. Then the masking layer is removed and after an appropriate washing an impurity is diffused or introduced in another way in the body of the silicon cell which was previously doped. The diffusion creates the usual junction n-p or p-n inside the cell surface, this junction extending over the entire surface of the cell including the indentations formed in this surface.

 A very simple and effective method is to use steam to form an oxide layer on a silicon wafer.

The vapor creates a layer of silicon dioxide which covers the entire main surface of the wafer to be attacked. By photolitography the layer of silicon dioxide is selectively attacked by an attack agent which easily attacks silicon dioxide but which does not easily attack silicon.

In this way a series of open regions is formed over the entire layer of silicon dioxide which covers the main surface of the cell. After these open regions have been formed, a silicon attack agent is applied to the open regions to form a series of inverted pyramids over the entire surface of the cell. The silicon dioxide layer is removed by attack in its entirety, the cell is washed and an impurity is diffused or applied in another way to the structured surface of the cell to form an electrical junction inside this surface .

Other advantages and characteristics of the invention will appear on reading the following description of an embodiment referring to the appended drawing in which
Fig. 1 is a diagram illustrating the succession of steps for implementing the method of the invention,
Fig. 2 is a top plan view of a surface of a solar cell obtained by implementing the method illustrated diagrammatically in FIG. 1.

 Referring to the drawing and in particular to the diagrams of FIG. 1 we see that we use a silicon wafer 10 of desired dimensions, for example a disc 7.62 cm in diameter and 0.254 mm thick composed of substantially monocrystalline silicon which has been doped with boron . Suitably protected, the wafer 10 which has a main surface 11 in the crystallographic plane 1-0-0 of the silicon is placed in a vapor atmosphere at 9000 C for half an hour. After steam treatment, the surface 11 of the cell is covered with a masking layer 12 of silicon dioxide. The layer 12 has a substantially constant thickness of 2,000 A.

A layer of a photoresist 14 is then uniformly applied to the coating or layer 12 of silicon dioxide. This layer of photoresist is for example that described in US patent application No. 614 618 filed on 18
September 1975, which illustrates the photolitographic techniques that can be used for the production of a patterned mask. The film layer 14 is then photographed in a pattern of tiny holes whose centers are spaced from each other by about 30 microns. A solvent is then applied and the parts of the film 14 which have been photographically exposed are dissolved in the solvent bath. With such dissolved parts, the silicon wafer has the shape shown in the third diagram in FIG. 1, that is to say that the parts 14 are separated from each other by open regions 16 at the top of the continuous layer of silicon dioxide 12.

 An etching agent is then applied to the whole of the wafer 10 for the silicon dioxide but not for the silicon or for the layer of photoresist. In the present case, the attacking agent used is hydrofluoric acid. After having been immersed in hydrofluoric acid, the plate substantially presents the shape illustrated in the fourth diagram of FIG. 1 that is, the open regions 16 in the photoresist layer 14 are extended through the silicon dioxide layer 12 down to the surface 11 of the cell. For clarity, these open regions thus extended have been designated by the reference numeral 17. The rest of the layer 14 is then removed by immersion in an organic solvent, for example acetone and the cell has the structure illustrated in the fifth diagram of FIG. 1, a structure in which the small open regions 18 always extend to the surface 11 of the silicon wafer 10. In this form there is a pattern of open regions 18 distributed uniformly over the entire main surface 11 of the brochure. This pattern is composed of a continuum of silicon dioxide 12 in which open regions 18 have been formed.

 During the next step, the wafer with its layer of silicon dioxide is immersed in a bath of silicon attack agent, in this case a solution. 5 of KOH at 70 to 800 C, for approximately 5 minutes. Because the KOH constitutes an attack agent for the silicon but does not attack in a satisfactory way making the silicon dioxide, after the attack by the KOH the layer 12 of silicon dioxide, perforated by the open regions 18 finds substantially unchanged. However, the attack on the surface 11 of the silicon wafer 10 has been accomplished. This attack in the 1-0-0 crystallographic plane of monocrystalline silicon has resulted in the attacked surface having inclined planes 19 which extend towards inside the body 10 of the wafer and terminate in vertices 20 extending downwards. The wafer is again subjected to treatment with hydrofluoric acid after which the silicon dioxide layer 12 is completely removed. The result obtained is a wafer as shown in the last diagram in FIG. 1 and comprising open zones or indentations 22 distributed uniformly in its main surface 11, the voids of these indentations being connected by inclined planes 19 ending in vertices 20.

We see in Fig. 2 a largely enlarged view of the cell schematically represented in the last diagram of the
Fig. 1. As we can see the attack by the KOH in the 1-0 plan
O of monocrystalline silicon has made it possible to obtain a surface with voids in the form of inverted pyramids having four inclined surfaces 19 ending in a single vertex 20 for each indentation or void. The indentations formed by the planes 19 cover more than 50% of the surface 11 remaining in the cell.

In the preferred embodiment, the indentations occupy more than 90% of the main surface 11 of the cell, the surface at which light is received and absorbed by the cell. According to the measurements that have been made, the top of each pyramidal indentation extends into the body of the cell to a depth of about 20 microns. The base of an inverted pyramidal void has a length and a width d '' about 15 microns; the width of the surface portion 11 of the cell surface remaining between the adjacent indentations and separating them is approximately 1 micron.

 According to an important characteristic of the present invention, the exposed areas of the surface 11 below the open regions in the coating 12 are substantially symmetrical. The use of the term substantially symmetrical does not exclude the exposed portions which, in plan view, would be other than round or square. Thus, within the framework of the present invention, open regions of which the ratio between the largest and the smallest dimension may be for example 2, although it is currently believed that the exposed perfectly symmetrical portions of the surface constitute the mode most preferable achievement.

 According to an additional characteristic of the invention, a structured surface of inverted pyramidal voids is produced on the two main surfaces, that is to say the front and rear surfaces of a silicon wafer. The voids on the back surface of the wafer, which should not be exposed to light, are then filled with a heat-conductive metal such as tin.

In this form the structured front structure of the cell performs its light absorption function and the rear surface acts as a heat collector to transfer heat away from the rest of the cell and then in contact with a heat dissipating medium , for example air or water.

 It should be understood that the process and the product which have just been illustrated and described represent only a preferred mode of implementation. Thus the specific dimensions of the inverted pyramidal indentations of the cell were used to illustrate the characteristics which are now considered to be the most advantageous. However, experimentation can show that other dimensions of indentation offer certain advantages, in particular in relation to the specific use for which the cell is intended. In addition methods other than photolitography can be used to arrange the open regions in silicon dioxide or another masking layer can be applied to the light impact surface of the cell. In addition such open regions can be otherwise defined, for example by applying a mask to the cell surface and attacking the indentations of the cell surface through the mask. However, since great precision is required and the mask itself must be resistant to the silicon attack agent, the method which has been described previously has proved to be particularly advantageous. Furthermore, the attack on the surface of the wafer can also take place through the open regions in a coating of titanium or titanium dioxide which would have the advantage of being applied at a lower temperature-avoiding possible thermal damage to the cell.

 Although the invention has been described in connection with a particular embodiment, it is in no way limited thereto and numerous variants and modifications can be made thereto without departing from its scope or its spirit.

Claims (13)

CLAIMS t. Method of manufacturing a silicon solar cell having at least one main surface adapted to receive light incident on it and to absorb and convert this light into electrical energy, characterized in that a silicon wafer is used which has at least one main surface capable of being attacked by a silicon attack agent, which this surface is masked with a layer resistant to attack by said silicon attack agent, said masking layer being provided with regions open that expose substantially symmetrical portions of said surface of the silicon wafer, as said exposed portions of said surface are attacked through said open regions of said masking layer with said silicon etchant to produce indentations in said surface, and that an electrical junction is formed on said surface.  2. A method of manufacturing a silicon solar cell according to claim 1, characterized in that said masking layer is substantially uniform and covers the whole of said main surface;  3. A method of manufacturing a silicon solar cell according to claim 1, characterized in that said masking layer is the reaction product of silicon and an externally applied reagent.  4. A method of manufacturing a silicon solar cell according to claim 3, characterized in that said open regions of said masking layer are formed by attack by an attack agent which does not easily attack silicon.  5. A method of manufacturing a silicon solar cell according to claim 1, characterized in that said masking layer consists of titanium.  6. A method of manufacturing a silicon solar cell having at least one main surface suitable for receiving and absorbing the light incident on it and for converting said light into electrical energy, characterized in that a silicon wafer is used having a main surface in the 1-0-0 plane of crystalline silicon, said silicon being subjected to attack by a silicon attack agent, which is reacted with an oxidizing agent to form a layer of a silicon oxide covering said surface, which is carried out a selective attack of said oxide layer on said surface using an agent not attacking the silicon to form open regions in said oxide layer passing through said oxide layer to expose portions lying below said main surface of said wafer, as said exposed portions of said surface are attacked through the open regions in said layer of oxide with a silicon attack agent which does not constitute an attack agent for said oxide layer to produce indentations in said main surface and that an electrical junction is formed on said surface.  7. Method for manufacturing a silicon solar cell according to claim 6, characterized in that said oxidizing agent is steam and that said oxide layer consists of silicon dioxide.  8. Method for manufacturing a silicon solar cell according to claim 6, characterized in that said oxide layer is removed after the formation of said indentations in said main surface.  9. Method for manufacturing a silicon solar cell according to claim 6, characterized in that # said oxide layer is selectively attacked through a coating. having spaced uncoated portions, said uncoated portions being produced by photolithography.  10. Silicon solar cell having a main surface adapted to receive the light incident on it, and to absorb and convert this light into electrical energy, characterized in that said surface has a plurality of indentations, said indentations being formed completely inside and delimited entirely by the body of said cell and having their bases in the plane of said main surface.  1.1. Solar silicon cell according to claim 10, characterized in that the said indentations are substantially in the form of pyramids, the bases of which lie in the plane of the said main surface.  12. Solar cell with silicon according to claim 11, characterized in that said pyramids have a substantially equal depth and are uniformly spaced along said main surface, the bases of said pyramids occupying more than 90% of said main surface.  13. Solar cell with silicon according to claim 10, characterized in that it has a rear surface opposite to said main surface also provided with said indentations extending inward.  14. Solar cell silicon according to claim 13, characterized in that the indentations of said rear surface are filled with a heat conductive metal.