FR2631330A1 - Glass with a transparent conductive layer for photopile in the form of a thin layer and process for obtaining it - Google Patents

Abstract

A glass with a transparent conductive layer intended to be used as electrode for photopiles in the form of a thin layer particularly those made of amorphous silicon. The process of preparation is also described. The layer consists of ITO obtained by the pyrolysis of a powder based on indium formate. The layer is made scattering by carrying out the coalescence of the grains of the layer. This is obtained by adding colloidal silica to the powder to be pyrolysed.

Description

TRANSPARENT CONDUCTIVE LAYER GLASS FOR PHOTOPILES
THIN FILM AND PROCESS FOR OBTAINING SAME
The invention relates to glasses covered with a transparent conductive layer intended to serve as an electrode for thin-film photovoltaic cells.

 One of the techniques used to make the electrode before thin-film solar cells consists in using a layer composed of transparent conductive metal oxides deposited on soda-lime-silica glass. Most often, use of tin oxides doped with halogens or indium oxides doped with tin.

 The requirements imposed on these layers relate to their electrical and optical characteristics and their behavior during the manufacture of the photovoltaic cell or during their use in operation.

 This behavior problem is particularly delicate when depositing a layer of hydrogen amorphous silicon (a-SiH) by the effluving technique of a silane gas. In this case, it can be seen that certain layers which are otherwise satisfactory, such as, for example, the layers of indium-tin oxides deposited by cathode sputtering are reduced in the plasma and rendered incapable of their function as transparent electrodes.

 But the other requirements are also essential: for power cells, such as those that operate in the sun to operate motors, the electrical resistance of the layer must be as low as possible so that the passage of high currents does not lower the voltage at the terminals too much. Optically, the first requirement concerns the thickness of the layer, which must be mastered for aesthetic reasons but above all to optimize the conversion efficiency into electrical energy / light energy. amorphous hydrated silicon (a-SiH) absorption in the photovoltaic material must be maximum for a wavelength of about 550 nm, which, taking into account the respective indices of the glass, of the transparent conductive layer itself and silicon, imposes a thickness for the conductive layer which it is imperative to respect. Still in the field of optics, we have long tried to increase the conversion efficiency mentioned above by lengthening the optical path of light in the photovoltaic layer by giving a diffusing character to the glass + transparent conductive layer To achieve this result, different methods have been proposed, most of which lead to the creation of a certain roughness at the interface between the transparent conductive layer and the photovoltaic material. Certain methods, such as that of patent application EP 106 540 or of patent application FR 2 578 359 propose to make the surface of the glass rough before the deposition of the conductive layer. Other techniques, as in the patent issued FR 2 514 201 propose that it is the layer itself which has a roughness.

 The invention has as its task to solve the following problem: obtaining a transparent conductive layer which has good conductivity, the thickness of which is easy to control and which can be made diffusing under economical industrial conditions.

 According to the invention, a glass is produced with a transparent electroconductive layer intended to constitute both the support and an electrode of an amorphous silicon photocell, characterized in that the layer comprises an indium oxide obtained by pyrolysis of a powder of an organometallic salt of indium suspended in a gas sprayed on the # hot glass.

 It is also proposed to use as base of the organometallic compound, indium formate. In a variant of the invention, the doping of indium is carried out with tin and at least in part with an organometallic compound also supplied in powder form.

The preferred embodiment of the invention uses dibutyltin oxide (DBTO) for this.

To make the layer diffusing, the invention proposes that its relief has a particle size whose average diameter is greater than 2
The proposed method for obtaining a transparent electroconductive layer diffusing light is characterized in that a powder is sprayed on the glass, suspended in a gas, based on an organometallic salt of indium added with an amount of at least 3% by weight of an anhydrous, hydrophobic product based on finely divided silica and preferably greater than 5 S.

 In a variant of the above process, the invention provides that the product contains more than 90% of silica.

 The preferred embodiment of the method uses the organometallic salt of indium, formate. Indium doping is preferably ~ rence performed with tin and at best, using dibutyltin oxide in powder form.

 In one embodiment of the invention, the method is used to obtain a transparent electroconductive layer diffusing light or a powder based on an organometallic salt d is sprayed onto the glass indium adds an amount of 5.5% by weight of an anhydrous, hydrophobic product based on finely divided silica. In another form, the process is used which consists in spraying onto the glass, using gas as a carrier, a powder based on an organometallic salt of indium added with an amount of 10% by weight of an anhydrous product. , hydrophobic based on finely divided silica.

 The techniques for distributing powders transported by a gas stream have already used finely divided colloTdale silica ~ in order to homogenize the powder. Thus the patent of ~ book EP 100 740 proposes to add to organometallic powders to be pyrolyzed on the glass quantities which can reach 5% of Siq but preferably limits to 2% because the higher quantities would cause difficulties.

 The invention will now be described in detail using the figures. Figure 1 shows the surface condition of a layer of indium tin oxide according to the invention. In FIG. 2, it is the surface condition of a layer of diffusing indium tin oxide that has been shown.

The installation shown diagrammatically in FIG. 3 made it possible to implement the invention. Figures 4 and 5 show the experimental results obtained with the diffusing layers according to the invention.

 As we have seen, for all power applications, it is necessary to have a very conductive layer as a transparent electrode. Among all those which could be envisaged, it is well known that it is the layers, either based on metallic silver or based on indium tin oxide (ITO) which have the best electrical performance. Unfortunately, in the case of hydrogenated amorphous silicon solar cells, silver - even protected by the dielectric layers usually used - does not resist very aggressive plasma at the temperature or it acts during the deposition of a-SiH. ITO are also known for their excellent electrical conductivity, which is in particular - equal thickness - greater than that of the layers of tin oxide doped with halogens. However, the use of ITO alone as a transparent photocell electrode has so far been unsuccessful. In fact, it can be seen during d8p8t by effluvation from silane gases that the usual ITO layers obtained by a CVD method, a spray method or else by cathode sputtering, ie from a metal target in the presence of oxygen-containing plasma, either from a target of indlum-etaln oxide, does not resist the aggression of the reducing plasma during the deposition of a-SiH. The stake is of importance since 'a large part of the thin film solar cells produced in the world is intended for energy uses, but until now If one wanted to use layers of ITO it was necessary to protect them by another layer, for example of oxide of tin which is more resistant to reducing plasma. But this solution is expensive because it requires an additional production phase, in addition, it has the disadvantage of adding a poorly conductive layer on top of the ITO, which degrades the electrical performance of the transparent electrode.

 Despite this state of affairs and given the importance of the issue, the applicant has carried out tests using the layers developed according to the invention described in patent application EP 192 009. It has been found that these layers, obtained by pyrolysis of a powder based on an organometallic compound of indium, were only very slightly degraded insofar as their temperature during deposition remained as low as possible and in no case exceeded 2500C.

 It is obvious that the application of the chemically resistant layers obtained by the process described in patent application EP 192 009 is not limited to thin-film photocells of the a-SiH type but also to other thin-film photocells when the transparent electrodes are subject to an aggressive atmosphere.

The goal that all producers of solar cells seek to achieve is to obtain the best possible electrical energy / light energy conversion yields. It is therefore a question of producing as many useful electrons as possible from the photons available. A
of the techniques employed concerns the transparent electrode which serves as the light entry face. We want to "trap" the light rays and - in the case of a-SiH - more particularly those of the red part of the spectrum. The method proposed here consists in giving the ITO layer described above a surface structure which diffuses the light.

 In Figure 1 we have shown the contours as they appear on a plate obtained by a scanning electron microscope.

This is the "normal" ITO that has been discussed so far. We see in the figure of grains 1 whose chemical composition is little different from that of the bottom of layer 2 which appears very regular.

The scale 3 at the edge of FIG. 1 represents 10 μm. It can therefore be seen that the grains have an average dimension of the order of a micron.

 In FIG. 2, represented on a scale where the segment 4 represents 10 μm, on the other hand, a diffusing layer according to the invention is reproduced, the contours of which - noted under the same conditions are represented. We see the same bottom layer 6. We also see, fewer, grains 5 of the same type as above but we especially notice areas 7 of much larger dimension. These zones have thicknesses of the same order as the grains 5, they appear to consist of the coalescence of these grains, the unitary structure of which is often visible in the zones 7. These zones give the layer of ITO performances which will be presented in detail later.

 Both to describe the way of producing the diffusing layers according to the invention and to compare their performance with the other solutions, we will describe the tests that have been carried out.

 Five samples were prepared using two different support glasses, on the one hand an extra-white glass (without iron in its composition) normally polished, of thickness 1.3 mm for four samples and a sample of an extra glass identical white but which has undergone a chemical matting treatment carried out in a known manner with a mixture of a commercial paste (LERITE brand, based on ammonium fluoride) and hydrochloric acid. The treatment is only done on one side and the surface obtained consists essentially of pyramids with a base 10 to 20 μm wide and a height between 2 and 5 μm.

 The samples were 70 x 60 mm in size.

 In a second operation, the chemically matte sample and a sample of the base glass were both subjected, separately to the ITO deposition treatment on a laboratory installation as shown in FIG. 3. The experimental procedure is the same in both cases: in an electric oven (not shown) located under the installation and heated to 600in, the glass 8 is held vertically for a period of 6 minutes. It then climbed back into its plane at a speed of 8 meters per minute. It then passes in front of a nozzle 9, the slot 10 of which has dimensions of 3 x 100 mm. The nozzle blows air at ordinary temperature with a flow rate of 37 Nm3 / h. A quantity of powder is then introduced into the air stream, at one time, using a funnel not shown. This powder consists of a mixture of indium formate for 90 X by weight, combined with dibutyltin oxide (DBTO, 10 X) the respective particle sizes of these two powders are 5 to 15 pm for the first and 7 at 23 fi for the second. The indium formate was prepared in the manner described in patent application EP 192 009. The quantity introduced is such that one obtains on each square centimeter of glass the projection of an amount of at least 16 grams of the mixture of powders. In general, these experimental conditions are sufficiently defined so that the desired thickness of ITO layer is immediately obtained, this should be such that, after the reducing heat treatment which will be discussed below , the layer- has a maximum of reflection at a wavelength of 500 nu (yellow of order 2). When this thickness was not obtained from the first test, slight modifications in the quantities of powder introduced made it possible to achieve the result.

 On the other three samples of polished extra-white glass was deposited diffusing ITO according to the invention. The method used is the same as before, the only difference lies in the nature of the powder introduced into the nozzle 9: it is here the same mixture as before but variable quantities of a cotordal silica of the AEROSIL type have been added marketed by DEGUSSA. The variety chosen is 12AEROSIL R 974 which has a specific surface of around 170 m2 / g and an average particle diameter of 12 nm. To allow a good coating of the grains of powders by the much finer particles, of finely divided silica, the powder is added in small increasing fractions in a receiver or a mixer containing all of the divided silica to be incorporated. Three different mixtures were thus produced: one with 1 X by weight of silica, and the other two with 5.5 and 10% respectively. The indium formate and the DBTO, which constitute the remainder of the mixture, are, for their part, in the same ratio with respect to each other as during the first deposits. The experimental procedure is also identical in every respect.

 The preparation of the samples is completed by a reduction annealing treatment. After cooling, the previous samples are taken up and brought to an oven where the temperature rises to 500 ° C., as soon as this temperature is reached, a nitrogen-hydrogen mixture is circulated in the proportions 95-5. The sample is kept for 10 minutes at this temperature and until the return to ordinary temperature which takes place gradually, the reducing atmosphere is maintained. A reducing annealing of the ITO layer is thus carried out, which increases its electrical conductivity.

 We thus have five different conductive substrates: a first "normal" ITO on clear glass, a second, "normal" ITO on mat glass, a third ITO diffusing at 1 X of SiO2 and the last two respectively at 5.5 and 10 % of SiO2.

 Figures 5 and 6 show the results obtained. The characteristics measured were in addition to the surface resistance Rz expressed in ohms, the blur H expresses in%, this value is the ratio between the diffuse transmission and the total transmission. This measurement is carried out here in monochromatic light, at a wavelength of 550 nm. These two measurements are made on a spectrophotometer equipped with an integrating sphere, without masking for the measurement of the total transmission and by eliminating the main beam for the diffuse transmission.

After these measurements were carried out on all the samples, these were used as a support for a deposition of a-SiH under the usual conditions. After the deposition, new optical measurements were made. These are total reflection measurements at a wavelength located in the red. The average of the reflection values is calculated in the 600-800 nm range and for each sample. The relative difference is then calculated: reflection of the reference sample (flat glass not treated) minus that of the sample studied, the everything being divided by the reflection of the reference sample1 it is therefore a relative value which is expressed in%, it is called ~ S
In FIG. 4, we see in 11, the value of H for the chemically mated sample, it is high: 82 X which is much higher than the values of blur obtained by the dtITO layer proper: as shown in the curve 12, this ranges from 3 X for 1 X of AEROSIL in the pyrolyzed powder on the glass during ITO deposition at 24%, maximum value reached for 10 X of AEROSIL. The value of the reference layer (OX of AEROSIL) is, with 2.8%, of the same order as with 1% of silica.

 Given these large differences in the light scattering before depositing a-SiH, one would have expected differences of the same order on which somehow measures the efficiency of trapping red light in the photovoltaic layer. The results of FIG. 4 show on the contrary a comparable effectiveness of the two techniques, chemical matting and ITO diffusing. With S = t 35%, the chemical matting at 13 has a value slightly higher than that (33%) reached by curve 14 for a quantity of SiO # of 10 X. With 29% the trapping efficiency is hardly less good at 5.5 X from AEROSIL. Tests of photovoltaic conversion efficiency proper will have to conform these first results.

 The surface resistance R # expressed in ohms is already shown in Figure 5. With 10 ohms, the resistance of the normal ITO layer on polished glass (O X of SiO2) is obviously the best. We see on curve 15 that with 1 X, of Siq we change practically nothing (11 t). On the other hand, we must note with the percentages of Si # effective in terms of diffusion, a slight degradation of the conduction of the layer since we reach values of 17 and 18 ohms per square. In comparison, the value represented in 16 of the resistance of the layer of ITO deposited on the chemical mat glass ~ ment remains lower (15 n).

 If we summarize the results obtained, we can say that the technique according to the invention allows, at the cost of a slight degradation of conductivity, to achieve a trapping efficiency of the same order as that of a chemically matte glass. The advantages of the technique described therefore lie in the economy of the chemical matting operation, an operation which is expensive and uses substances based on hydrofluoric acid, dangerous for operators as well as for the environment.

 More generally, the use of a layer of ITO pyrolyzed from a powder based on indium formate has the advantage of very good electrical conduction, of excellent regularity. thickness and lower production cost than diapers. identical performance, for example ITO by sputtering covered with SnO2 obtained by the same technique.

Claims (13)

 1. Glass with transparent electroconductive layer intended to constitute both the support and an electrode of an amorphous silicon solar cell, characterized in that the layer comprises an indium oxide obtained by pyrolysis of a powder of an organometallic salt indium suspended in a gas sprayed on the hot glass.  2. Layered glass according to claim 1, characterized in that the organometallic salt is based on indium formate.  3. Layered glass according to claim 1, characterized in that the doping of indium oxide is carried out with tin.  4. Layered glass according to claim 3, characterized in that the tin is provided in pulverulent form based on a ganometallic gold compound of tin.  5. Layered glass according to claim 4, characterized in that the organometallic tin compound is dibutyltin oxide (DBTO).  6. Glass with transparent electroconductive layer according to claim 1, characterized in that the layer is made diffusing by the constitution of a relief whose grains have an average diameter greater than 2 µm.  7. Method # for obtaining a transparent electroconductive layer diffusing light, characterized in that a powder based on a ganometallic gold salt of indium added is sprayed onto the glass, suspended in a gas of an amount of at least 3 X by weight of an anhydrous, hydrophobic product based on finely divided silica.  8. Method according to claim 7, characterized in that this product contains more than 90 X of silica.  9. Method according to claim 7, characterized in that the organometallic salt is based on indium formate.  10. Method according to claim 9, characterized in that the doping of indium oxide is carried out with tin.  11. Method according to claim 10, characterized in that added to the indium formate powder, dibutyltin oxide powder.  12. A process for obtaining a transparent electroconductive layer diffusing light, characterized in that a powder based on an organometallic indium salt added with an amount of 5.5% by weight of an anhydrous, hydrophobic product based on finely divided silica.  13. Process for obtaining a transparent electroconductive layer diffusing light, characterized in that a powder based on a ganometallic or indium gold salt with added d is sprayed onto the glass, suspended in a gas an amount of 10% by weight of an anhydrous, hydrophobic product based on finely divided silica.