Process for the production of solar cells using printable doping media which inhibit the diffusion of phosphorus

Abstract

The present invention relates to a novel printable medium in the form of a hybrid sol and/or gel based on precursors of inorganic oxides for use in a simplified process for the production of solar cells, in which the medium according to the invention functions both as doping medium and also as diffusion barrier.

Claims

1. Printable hybrid sols and/or gels based on precursors of inorganic oxides which are printed selectively onto suitable surfaces of the substrate by means of suitable printing processes on silicon surfaces for the purposes of local and/or full-area diffusion and doping on one side for the production of solar cells, dried and subsequently brought to specific doping of the substrate itself by means of a suitable high-temperature process for release of the boron oxide precursor present in the hybrid gel to the substrate located beneath the hybrid gel.

2. Hybrid sols and/or gels according to claim 1, characterised in that they are compositions based on precursors of silicon dioxide, aluminium oxide and boron oxide.

3. Hybrid sols and/or gels according to claim 1, characterised in that they are compositions based on precursors of silicon dioxide, aluminium oxide and boron oxide which are employed as a mixture.

4. Printable hybrid sols and/or gels according to claim 1, characterised in that they have been obtained on the basis of precursors of silicon dioxide, selected from the group of symmetrically or asymmetrically mono- to tetrasubstituted carboxy-, alkoxy- and alkoxyalkylsilanes, in particular alkylalkoxysilanes in which at least one hydrogen atom is bonded to the central silicon atom, carboxy-, alkoxy- and alkoxyalkylsilanes, in particular alkylalkoxysilanes, which contain individual or different saturated, unsaturated branched, unbranched aliphatic, alicyclic and aromatic radicals, which may in turn be functionalised at any desired position of the alkyl, alkoxide or carboxyl radical by heteroatoms selected from the group O, N, S, Cl and Br, and mixtures of these precursors.

5. Printable hybrid sols and/or gels according to claim 1, characterised in that they have been obtained on the basis of precursors of silicon dioxide, selected from the group triethoxysilane, tetraethyl orthosilicate, triethoxysilane, ethoxytrimethylsilane, dimethyldimethoxysilane, dimethyldiethoxysilane, triethoxyvinylsilane, bis[triethoxysilyl]ethane and bis[diethoxymethylsilyl]ethane, and mixtures thereof.

6. Printable hybrid sols and/or gels according to claim 1, characterised in that they have been obtained on the basis of precursors of aluminium oxide, selected from the group of symmetrically and asymmetrically substituted aluminium alcoholates (alkoxides), aluminium tris(-diketones), aluminium tris(-ketoesters), aluminium soaps, aluminium carboxylates, and mixtures thereof.

7. Printable hybrid sols and/or gels according to claim 1, characterised in that they have been obtained on the basis of precursors of aluminium oxide, selected from the group aluminium triethanolate, aluminium triisopropylate, aluminium tri-sec-butylate, aluminium tributylate, aluminium triamylate and aluminium triisopentanolate, aluminium acetylacetonate or aluminium tris(1,3-cyclohexanedionate), aluminium mono-acetylacetonate monoalcoholate, aluminium tris(hydroxyquinolate), mono- and dibasic aluminium stearate and aluminium tristearate, aluminium acetate, aluminium triacetate, basic aluminium formate, aluminium triformate and aluminium trioctanoate, aluminium hydroxide, aluminium metahydroxide and aluminium trichloride, and mixtures thereof.

8. Printable hybrid sols and/or gels according to claim 1, characterised in that they have been obtained on the basis of precursors of boron oxide, selected from the group of alkyl borates, boric acid esters of functionalised 1,2-glycols, boric acid esters of alkanolamines, mixed anhydrides of boric acid and carboxylic acids, and mixtures thereof.

9. Printable hybrid sols and/or gels according to claim 1, characterised in that they have been obtained on the basis of precursors of boron oxide, selected from the group boron oxide, diboron oxide, triethyl borate, triisopropyl borate, boric acid glycol ester, boric acid ethylene glycol ester, boric acid glycerol ester, boric acid ester of 2,3-dihydroxysuccinic acid, tetraacetoxy diborate, and boric acid esters of the alkanolamines ethanolamine, diethanolamine, triethanolamine, propanolamine, dipropanolamine and tripropanolamine.

10. Printable hybrid sols and/or gels obtainable by bringing precursors of claim 4 to partial or complete intra- and/or interspecies condensation under water-containing or anhydrous conditions with the aid of the sol-gel technique, either simultaneously or sequentially, forming storage-stable, very readily printable and printing-stable formulations.

11. Printable hybrid sols and/or gels according to claim 10, obtainable by removal of the volatile reaction assistants and by-products during the condensation reaction.

12. Printable hybrid sols and/or gels according to claim 10, obtainable by adjustment of the precursor concentrations, the water and catalyst content and the reaction temperature and time.

13. Printable hybrid sols and/or gels according to claim 10, obtainable by specific addition of condensation-controlling agents in the form of complexing agents and/or chelating agents, various solvents in defined amounts, based on the total volume, whereby the degree of gelling of the hybrid sols and gels formed is specifically controlled.

14. Use of the printable hybrid sols and/or gels according to claim 1 in a process for the production of solar cells, in which they are processed and deposited by means of a printing process selected from spin or dip coating, drop casting, curtain or slot-die coating, screen or flexographic printing, gravure, ink-jet or aerosol-jet printing, offset printing, microcontact printing, electrohydrodynamic dispensing, roller or spray coating, ultrasound spray coating, pipe-jet printing, laser transfer printing, pad printing, flat-bed screen printing and rotary screen printing.

15. Use of the printable hybrid sols and/or gels according to claim 1 for the processing of silicon wafers for photovoltaic, microelectronic, micromechanical and micro-optical applications.

16. Use of the printable hybrid sols and/or gels according to claim 1 for the production of PERC, PERL, PERT and IBC solar cells and others, where the solar cells have further architectural features, such as MWT, EWT, selective emitter, selective front surface field, selective back surface field and bifaciality.

17. Use of the printable hybrid sols and/or gels according to claim 1 as boron-containing doping medium for silicon, where the medium simultaneously acts as diffusion barrier or as diffusion-inhibiting layer against undesired diffusion of phosphorus through this medium and completely blocks or sufficiently inhibits the latter so that the doping prevailing beneath these printed-on media is p-type, i.e. boron-containing.

18. Use according to claim 17, characterised in that doping of the printed substrate is carried out by suitable temperature treatment and doping of the unprinted silicon wafer surfaces with dopants of the opposite polarity is induced simultaneously and/or sequentially by means of conventional gas-phase diffusion, where the printed-on hybrid sols and/or gels act as diffusion barrier against the dopants of the opposite polarity.

19. Process for the doping of silicon wafers, characterised in that a) silicon wafers are printed locally on one or both sides or over the entire surface on one side with the hybrid sols and/or gels according to claim 1, the printed-on medium is dried, compacted and subsequently subjected to subsequent gas-phase diffusion with, for example, phosphoryl chloride, giving p-type dopings in the printed regions and n-type dopings in the regions subjected exclusively to gas-phase diffusion, or b) hybrid sol and/or gel according to claim 1 deposited over a large area on the silicon wafer is compacted and local doping of the underlying substrate material is initiated from the dried and/or compacted medium with the aid of laser irradiation, followed by high-temperature treatment, inducing diffusion and doping for the production of two-stage p-type doping levels in the silicon, or c) the silicon wafer is printed locally on one side with hybrid sols and/or gels according to claim 1, where the structured deposition may optionally have alternating lines, the printed structures are dried and compacted, and the silicon wafer is subsequently coated over the entire surface on the same side of the wafer with the aid of PVD- and/or CVD-deposited phosphorus-doping dopant sources, where the printed structures of the hybrid sols and/or gels are encapsulated, and the entire overlapping structure is brought to structured doping of the silicon wafer by suitable high-temperature treatment, where the printed-on hybrid gel acts as diffusion barrier against the phosphorus-containing dopant source located on top and the dopant present therein, or d) the silicon wafer is printed locally on one side with hybrid sols and/or gels according to claim 1, where the structured deposition may optionally have alternating lines, the printed structures are dried and compacted, and the silicon wafer is subsequently coated over the entire surface on the same side of the wafer with the aid of doping inks or doping pastes which have a phosphorus-doping action, where the printed structures of the hybrid sols and gels are encapsulated, and the entire overlapping structure is brought to structured doping of the silicon wafer by suitable high-temperature treatment, where the printed-on hybrid gel acts as diffusion barrier against the phosphorus-containing dopant source located on top and the dopant present therein.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0057] Surprisingly, it has been found that printable hybrid sols and hybrid gels which consist at least of the following oxide precursors aluminium oxide, silicon dioxide and boron oxide are suitable as printable doping media for the local doping of silicon wafers and at the same time allow the phosphorus diffusion of the same wafers printed with these hybrid sols and gels, where the printed hybrid sols and gels act as efficient diffusion barrier against phosphorus diffusion. In other words, exclusive doping with boron is obtained under the co-diffusion conditions outlined in the regions printed with the sols and gels according to the invention, and exclusive doping with phosphorus is obtained in the regions exposed to the phosphorus oxide vapour having a doping action. The hybrid sols and gels according to the invention are described in the following documents: WO 2012/119686 A, WO2012119685 A1, WO2012119684 A, EP12703458.5 and EP12704232.3, and these should thus be regarded as part of the present disclosure.

[0058] The use of the hybrid sols and gels according to the invention thus enables simplified production of either solar cells which have structured dopings, such as, for example, IBC cells, or very generally of cells which have at least two different, not necessarily opposite dopings. Possible uses of the doping media according to the invention are outlined below.

[0059] FIG. 2 shows a simplified process flow chart for the production of a bifacial (in this case n-type) solar cell (PERT structure). The dopant source for the phosphorus diffusion was assumed to be a phosphorus paste, but it can equally well be any other source deposited over the entire surface, such as, for example, a doping ink or a CVD glass, a sputtered-on layer, epitactically deposited phosphorus-doped silicon, or a phosphorus-enriched silicon nitride layer. Instead of the boron paste according to the invention mentioned in the figure, it is of course also possible to use a boron doping ink according to the invention.

[0060] The production of a bifacial cell in accordance with FIG. 2 comprises the following essential process steps: two printing steps for the printing of the wafer surface, driving-in of the dopants, removal of the glass. A total of 4 process steps, compared with at least six, two of which are high-temperature steps, on use of the classical gas-phase diffusion process: masking on one side, diffusion with B, removal of the oxides, masking of the surface that has already been doped, diffusion with P, removal of the oxides. The use of the boron-containing doping media according to the invention results in a nominal reduction of the process steps necessary by one third compared with the classical process variant, which can thus be translated into more favourable production costs. The bifacial cell shown above can also be produced by the use of other dopant sources which can be deposited on one side, such as, for example, phosphorus-containing doping inks or a CVD glass, a sputtered-on layer, epitactically deposited and phosphorus-doped silicon, or a phosphorus-enriched silicon nitride layer, where the reverse procedure is of course also conceivable in principle in this connection. If the above-mentioned dopant sources acting on one side are used, a shortening of the process sequence can likewise be achieved: deposition of source 1, deposition of source 2, diffusion, removal of the residues. If both sources can be removed in one etching step, nominally the same effort as in the case already explained above arises: four process steps. This even applies in the case if the boron doping source according to the invention, ink (hybrid sol) or paste (hybrid gel), were to be replaced by one of the sources mentioned above, such as, for example, a CVD glass. However, the deposition of CVD glass as a vacuum process is a fairly expensive process step owing to the vacuum conditions. The same also applies to sputtering or epitactic deposition, meaning that the use of the boron doping media according to the invention has an inherent cost advantage owing to the less expensive deposition ability by means of printing steps. In principle, the wafer surface printed on both sides with doping media represents the least expensive possibility. In accordance with the composition according to the invention of the boron-containing hybrid sols and gels, parasitic dopings, which frequently take place and are to be observed from phosphorus-containing doping media, also do not represent a significant restriction of the possibility for the production of a bifacial solar cell by this route: the boron-containing hybrid sols and gels according to the invention, besides their function as dopant source, act as diffusion barriers for phosphorus diffusions. Surprisingly, it has therefore been observed that bifacial solar cells can be produced in a simple manner with the aid of the hybrid sols and gels according to the invention in accordance with the scheme outlined in FIG. 3.

[0061] FIG. 3 shows a simplified process flow chart for the production of a bifacial (in this case n-type) solar cell (PERT structure). A co-diffusion process using classical diffusion with phosphoryl chloride is depicted. The hybrid sols and gels according to the invention (here only mentioned as boron paste), besides their function as dopant source, act as diffusion barrier against phosphorus diffusion.

[0062] The hybrid sols and gels according to the invention act as diffusion barrier against phosphorus diffusion and thus protect the silicon wafer against penetration of this dopant from the gas phase into the surface regions of the wafer. At the same time, the boron-doping action of the hybrid sols and gels according to the invention is retained and thus enables on the one hand protection against penetration of phosphorus into the semiconductor and on the other hand effective diffusion and doping of the surfaces printed with these media with the desired and intended boron doping. Performance of the production of a bifacial solar cell in accordance with the principle outlined above results in the following essential process steps: printing of the boron source, co-diffusion with a phosphorus source from the gas phase, removal of the oxides and glassesin total three process steps. This thus corresponds to a reduction of the process steps necessary for the production of a bifacial solar cell by half compared with the classical procedure (gas-phase diffusion with masking), and a reduction by a quarter of the process steps necessary compared with the case outlined above using, for example, CVD-based or similar dopant sources. Co-diffusion with the boron-containing hybrid sols and gels according to the invention as dopant sources thus represents the least expensive possibility for the production of bifacial solar cells. It goes without saying in this connection that, with inclusion of European Patent Applications 14004453.8 and 14004454.6, selectively doped structures, at least in the regions to be doped with boron, of the wafer can also be produced very simply (cf. FIG. 4).

[0063] FIG. 4 shows a simplified diagrammatic representation of a bifacial solar cell (n-type) with selective or two-stage doping (selective boron emitter) in the region of the boron emitter.

[0064] The following figures depict the process sequences already outlined for the bifacial n-type solar cells. FIGS. 5 to 7 show the possible process sequences and the results thereof for p-type wafers as base material. The conclusions which can basically be derived for these process sequences are the same as already stated for the case of the production of bifacial n-type cells.

[0065] FIG. 5 shows a simplified process flow chart for the production of a possible bifacial (in this case p-type) solar cell (PERT structure). The dopant source for the phosphorus diffusion was assumed to be a phosphorus paste, but it can equally well be any other source deposited over the entire surface, such as, for example, a doping ink or a CVD glass, a sputtered-on layer, epitactically deposited phosphorus-doped silicon, or a phosphorus-enriched silicon nitride layer. Instead of the boron paste according to the invention mentioned in the figure, it is of course also possible to use a boron doping ink according to the invention.

[0066] FIG. 6 shows a simplified process flow chart for the production of a possible bifacial (in this case p-type) solar cell (PERT structure). A co-diffusion process using classical diffusion with phosphoryl chloride is depicted. The hybrid sols and gels according to the invention (here only mentioned as boron paste), besides their function as dopant source, act as diffusion barrier against phosphorus diffusion.

[0067] FIG. 7 shows a simplified diagrammatic representation of a possible bifacial solar cell (p-type) with selective or two-stage doping (selective back surface field) in the region of the boron back surface field.

[0068] FIG. 8 shows a possible process sequence for the production of a p-type PERL solar cell. The representation outlined in the scheme is based on the use of dopant sources which can be deposited over the entire surface and in a structured manner. In the case of the use of classical gas-phase diffusion, such as, for example, as a consequence of the use of phosphoryl chloride for phosphorus diffusion, an additional mask step for protection of the back-surface, open and base-doped regions of the wafer would also be necessary.

[0069] FIG. 8 shows a simplified diagrammatic representation of a possible production process of a p-type solar cell with back surface local contacts (PERL structure). The dopant source for the phosphorus diffusion was assumed to be a phosphorus paste, phosphorus ink or a CVD glass, but it can equally well be any other source deposited over the entire surface, such as, for example, a sputtered-on layer, epitactically deposited phosphorus-doped silicon, or a phosphorus-enriched silicon nitride layer. Instead of the boron paste according to the invention mentioned in the figure, it is of course also possible to use a boron doping ink according to the invention.

[0070] Let us now turn to the production of an IBC solar cell. With the classical procedure, based on gas-phase diffusion, we have seen that a sequence consisting of nine process steps is necessary in order to achieve the structured doping regions. FIG. 9 shows an alternative procedure which is based on the use of doping media to be applied in a structured manner to the back surface, such as the hybrid sols and gels according to the invention, while a further doping source can be applied over the entire surface to the front surface of the wafer. The doping source on the front side may likewise be, but does not necessarily have to be, a doping medium according to the invention, hybrid sol and/or hybrid gel. The alternative dopant sources already mentioned, such as, for example, a CVD glass, are likewise suitable. The above-mentioned naturally applies to the back surface. If the production process based on the hybrid sols and gels according to the invention is considered, five process steps are necessary in order to achieve a structured doping: deposition of source 1, deposition of source 2, deposition 3, high-temperature co-diffusion of all sources in a conventional tubular oven, removal of the dopant sources. This thus corresponds to a nominal reduction of the process steps necessary by 45% compared with the use of the classical diffusion process, which thus has the effect of significant advantages in the process costs. The same result is obtained on use of, for example, the deposition of CVD glasses as dopant source over the entire front surface. Since the use previously required deposition of the dopant source by a vacuum process, the achievable cost savings are not of the same order of magnitude as obtained in the case of the use of the printable hybrid sols and gels according to the invention as dopant source. If the CVD glasses having a doping action are also used for the definition of the doped regions present on the back surface, the requisite masking and structuring processes mean additional use of at least one structuring and etching step (previously counted as one unit); to this extent, the deposition of a further capping layer in between which separates two doped CVD glasses deposited one on top of the other can be omitted. Compared with the procedure based on gas-phase diffusion, a reduction of the process steps necessary by one third would thus be achieved. Compared with the use of the hybrid sols and gels according to the invention having a doping action, by contrast, additional use of a further process step or one fifth would arise. It goes without saying that a process sequence based on printing of the hybrid sols and gels according to the invention is preferable to the other process sequences outlined from the point of view of costs.

[0071] FIG. 9 shows a simplified diagrammatic representation of a possible production process of an n-type IBC solar cell. The dopant source for the front surface phosphorus diffusion was assumed to be a phosphorus paste, phosphorus ink or a CVD glass, but it can equally well be any other source deposited over the entire surface, such as, for example, a sputtered-on layer, epitactically deposited phosphorus-doped silicon or a phosphorus-enriched silicon nitride layer. On the back surface, the structured diffusion is obtained with the aid of various doping media, in this case referred to as boron and phosphorus paste. Instead of the boron paste according to the invention mentioned in the figure, it is of course entirely freely possible to use a boron doping ink according to the invention.

[0072] A further simplification of the production of IBC solar cells arises from the process flow chart depicted diagrammatically in FIG. 10. In this process flow chart, the property of the hybrid sols and gels according to the invention to act as diffusion barrier for phosphorus diffusion is utilised thoroughly. Consequently, as in the above-mentioned example, five process steps are used in order to achieve structured doping for IBC cells (in this case including the front surface doping, which is not depicted in the figure): deposition of source 01, deposition of source 2, deposition of source 3, high-temperature co-diffusion of all doping sources, removal of the dopant sources. In this example, a further cost reduction can be achieved compared with the example outlined above by depositing the back surface phosphorus source, for example, as doping ink using a very high-throughput deposition step. Such a step is, for example, the spray coating of the entire wafer surface. Alternatively, it may also be a flexographic printing step, which is claimed to have up to 2.5 to 3.0 times the wafer throughput compared with a conventional screen printing line. If the boron-containing doping source according to the invention is likewise deposited on the wafer surface, further cost reduction potentials can also be exploited compared with fairly inexpensive processing by screen printing.

[0073] FIG. 10 shows a simplified diagrammatic representation of a possible production process of an n-type IBC solar cell. Diffusion of the front surface was not considered in this case. The structured diffusion is achieved on the back surface with the aid of various doping media, in this case the structured application of the boron paste according to the invention, which can in principle equally well be a boron ink according to the invention. The back surface of the wafer is subsequently coated over the entire surface with a further phosphorus-containing dopant source, where the printed-on boron-containing doping medium according to the invention is likewise covered by the phosphorus-containing source. In the region not coated with the boron-containing medium according to the invention having a doping action, the phosphorus-containing source lies directly on the wafer surface and is able to dope this correspondingly with phosphorus during a high-temperature process, whereas in the regions of the boron-containing doping medium according to the invention, this acts as diffusion barrier against phosphorus diffusion and thus protects the wafer surface against penetration of phosphorus, but is at the same time capable of releasing the dopant, in this case boron, present in the medium to the wafer and thus inducing doping thereof with boron. Structured p/n junctions with an alternating sequence of the various doping regimes arise. The CVD doping glass mentioned in the figure can easily be replaced here by alternative dopant sources, such as, for example, a doping ink, a sputtered-on layer, epitactically deposited phosphorus-doped silicon, or a phosphorus-enriched silicon nitride layer.

[0074] The production of an IBC solar cell can furthermore be simplified by rigorous utilisation of the diffusion barrier properties of the hybrid sols and gels according to the invention against phosphorus diffusion. In this simplification, use is made of a co-diffusion step for obtaining boron doping with simultaneous or consecutive diffusion with phosphorus owing to the, for example, thermal decomposition of phosphoryl chloride. Both features are carried out in a single process step in a conventional tubular oven process. The wafer is subsequently treated on the front surface by means of one-sided etching in such a way that the front surface doping is adjusted to a certain, desired measure of the sheet resistance (cf. FIG. 11). The latter is necessary since IBC solar cells generally have weaker doping on the front surface than at the back surface contact points, the local back surface field. Lower doping on the front surface promotes the passivation capacity of this surface, which is accompanied by a reduction in the dark current saturation density and thus an increase in the cell voltage. The latter is ultimately evident from an increase in efficiency or as one of the most important levers for influencing the efficiency of a solar cell, in this case positively. As a consequence, the following number of steps arises as a process chain for obtaining structured dopings: deposition of the boron source, high-temperature diffusion in the presence of a reactive phosphorus precursor, back-etching of the front surface doping and removal of the dopant sources. In summary, these are four process steps. If we now combine the etching of the front surface field and the removal of the glass as one process step, which could be regarded as justified inasmuch as this practice was likewise used in the listing of the requisite process steps in the course of the classical procedure for obtaining structured doping (cf. structuring and etching), but much more decisively since this process sequence already was or in part still is established in industrial manufacture in this form for the production of selective emitter solar cells, then in total three process steps arise. As a consequence, the production of IBC solar cells with the aid of the hybrid sols and gels according to the invention thus opens up the possibility of saving six process steps or two-thirds of the requisite effort compared with the classical, purely gas phase-promoted doping. In principle, the same process chain can also be achieved with alternative conventional PVD- or CVD-deposited dopant sources. However, these must be structured on the back surface after deposition in order to define regions which are to be doped by means of the gas-phase processin this case practically likewise with the aid of phosphorus diffusion with phosphoryl chloride. The process chain arising from this thus necessarily has four process steps. Furthermore, a capping layer generally has to be incorporated in order to suppress penetration of the BSG glass applied in this case by the phosphorus diffusion with diffusing phosphorus. It thus becomes apparent that the use of the hybrid sols and gels according to the invention, which have a doping action and also act as barrier to phosphorus, has an inherent advantage which can significantly contribute to the cost-efficient production of IBC solar cells.

[0075] FIG. 11 shows a simplified diagrammatic representation of a possible production process of an n-type IBC solar cell. Diffusion of the front surface was considered in this case. On the back surface, the structured diffusion is achieved with the aid of various doping media, in this case the structured application of the boron paste according to the invention, where it can in principle equally well be a boron ink according to the invention. The wafer is subsequently subjected to conventional gas-phase diffusion with, for example, phosphoryl chloride as dopant precursor. All open points of the silicon wafer are thereby doped with phosphorus. The areas on the back surface which have been printed with the boron-containing dopant according to the invention are, owing to its property of acting as diffusion barrier against phosphorus diffusion, not doped with phosphorus, but instead by the boron present in the dopant source. The desired structured doping is consequently obtained on the back surface. The front surface may have been, but does not necessarily have to have been, subjected to excessive doping in the process. The doping intensity of the front surface is adjusted specifically against the desired requirements by controlled back-etching of the regions doped the most.

[0076] In the following examples, the preferred embodiments of the present invention are reproduced.

[0077] As stated above, the present description enables the person skilled in the art to use the invention comprehensively. Even without further comments, it will therefore be assumed that a person skilled in the art will be able to utilise the above description in the broadest scope.

[0078] Should anything be unclear, it goes without saying that the cited publications and patent literature should be consulted. Accordingly, these documents are regarded as part of the disclosure content of the present description. This applies in particular to the disclosure contents of the European patent applications with the file references 14004453.8 and 14004454.6 and the international application WO 2014/101990 A.

[0079] For better understanding and in order to illustrate the invention, examples are given below which are within the scope of protection of the present invention. These examples also serve to illustrate possible variants. Owing to the general validity of the inventive principle described, however, the examples are not suitable for reducing the scope of protection of the present invention to these alone.

[0080] Furthermore, it goes without saying to the person skilled in the art that, both in the examples given and also in the remainder of the description, the component amounts present in the compositions always only add up to 100% by weight, mol-% or vol.-%, based on the entire composition, and cannot exceed this, even if higher values could arise from the percent ranges indicated. Unless indicated otherwise, % data are therefore regarded as % by weight, mol-% or vol.-%.

[0081] The temperatures given in the examples and description and in the claims are always in C.

EXAMPLES

Example 1

[0082] 55.2 g of ethylene glycol monobutyl ether (EGB) and 20.1 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 7.51 g of glacial acetic acid, 0.8 g of acetaldoxime and 0.49 g of acetylacetone are added to this mixture with stirring.

[0083] 1.45 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for five hours. After warming, the mixture is subjected to a vacuum distillation at 70 C. until a final pressure of 30 mbar has been reached. The mass loss of readily volatile reaction products is 12.18 g. The distilled mixture is subsequently diluted with 62.3 g of Texanol and a further 65 g of EGB, and a mixed condensed sol consisting of precursors of boron oxide and silicon dioxide is added. The hybrid sol comprising silicon dioxide and boron oxide is to this end prepared as follows: 6.3 g of tetraacetoxy diborate are initially introduced in 40 g of benzyl benzoate, and 15 g of acetic anhydride are added. The mixture is warmed to 80 C. in an oil bath, and, when a clear solution has formed, 4.6 g of dimethyldimethoxysilane are added to this solution, and the entire mixture is left to react for 45 minutes with stirring. The hybrid sol is subsequently likewise subjected to a vacuum distillation at 70 C. until a final pressure of 30 mbar has been reached, where the mass loss of readily volatile reaction products is 7.89 g. 9 g of Synchro wax are added to the entire 110 g of mixture, and the mixture is warmed at 150 C. with stirring until everything has dissolved and the mixture is clear. The mixture is subsequently allowed to cool with vigorous stirring. A pseudoplastic and very readily printable paste forms.

Example 2

[0084] The paste according to Example 1 is printed onto a wafer with the aid of a conventional screen-printing machine and a 350 mesh screen with a wire thickness of 16 m (stainless steel) and an emulsion thickness of 8-12 m using a doctor-blade speed of 170 mm/s and a doctor-blade pressure of 1 bar and subsequently subjected to drying in a through-flow oven. The heating zones in the through-flow oven are for this purpose set to 350/350/375/375/375/400/400 C.

[0085] FIG. 12 shows a silicon wafer printed with the aid of the hybrid gel according to the invention in accordance with the composition and preparation of Example 1 after drying in a through-flow oven.

Example 4

[0086] The paste according to Example 1 is printed over a large area onto a rough CZ wafer surface (n-type) with the aid of a conventional screen-printing machine and a 280 mesh screen with a wire thickness of 25 m (stainless steel). The wet application rate is 1.5 mg/cm.sup.2. The printed wafer is subsequently dried at 300 C. on a conventional laboratory hotplate for 3 minutes and subsequently subjected to a diffusion process. To this end, the wafer is introduced into a diffusion oven at approximately 700 C., and the oven is subsequently heated to a diffusion temperature of 950 C. The wafer is kept at this plateau temperature for 30 minutes in a nitrogen atmosphere comprising 0.2% v/v of oxygen. After the boron diffusion, the wafer is subjected to phosphorus diffusion with phosphoryl chloride at low temperature, 880 C., in the same process tube. After the diffusions and cooling of the wafer, the latter is freed from glasses present on the wafer surfaces by means of etching with dilute hydrofluoric acid. The region which had previously been printed with the boron paste according to the invention has a hydrophilic wetting behaviour on rinsing of the wafer surface with water, which represents a clear indication of the presence of a boron skin in this region. The sheet resistance determined in the surface region printed with the boron paste is 195 /sqr (p-type doping). The regions not protected by the boron paste have a sheet resistance of 90 /sqr (n-type doping). The SIMS (secondary ion mass spectrometry) depth profile of the dopants is determined in the region of the surface which was printed by means of the boron paste according to the invention. In the region covered with the B paste, boron doping extending from the wafer surface into that of the silicon is determined, apart from the n-type base doping. The printed-on paste layer thus acts as diffusion barrier against typical phosphorus diffusion.

[0087] FIG. 13 shows the SIMS profile of a rough silicon surface which has been printed with the boron paste according to the invention and subsequently subjected to gas-phase diffusion with phosphoryl chloride. Owing to the rough surface, only relative concentrations in the form of count rates can be obtained.

Example 4

[0088] 3.66 g of boric acid which has been pre-dried in a desiccator are dissolved in 40 g of dibenzyl ether, 8 g of acetic anhydride and 8 g of tetraethyl orthosilicate in a round-bottomed flask at 100 C. with stirring and left to react for 30 minutes. 60 g of ethylene glycol monobutyl ether are subsequently dissolved in a solution of 0.4 g of 1,3-cyclohexanedione and 7.2 g of salicylic acid, and 160 g of ethanol are added. When the reaction mixture has mixed completely, 20 g of aluminium tri-sec-butylate are added to this solution. The solution is refluxed for a further hour. The boron ink is subsequently filtered through a filter having a pore size of 0.45 m and deaerated. For printing by means of an ink-jet printer, the ink is introduced into a suitable print head, Spectra SE128AA, and printed onto silicon wafers which have been subjected to acidic polish-etching with selection of the following printing conditions: firing frequency1500 Hz; voltage70 V; trapezium function1-111 s; reduced-pressure difference above the ink tank21.5 mbar. The substrates are warmed from below on the substrate holder. The respective warming (.fwdarw.printing temperature) is mentioned in the examples given. Squares having an edge length of 2 cm each are printed onto the wafers. The selected print resolution is likewise reproduced in the individual examples. After the printing, the printed wafers are dried at temperatures between 400 C. and 600 C. on a conventional laboratory hotplate for five to ten minutes in each case. The dried structures are subsequently printed with a phosphorus ink, in accordance with the composition as mentioned in the patent application WO 2014/101990, likewise by means of ink-jet printing. The respective print resolution, and also the respective printing temperature, is reproduced in the examples. The phosphorus ink is processed with selection of the following printing conditions: firing frequency1500 Hz; voltage90 V; trapezium function1-111 s; reduced-pressure difference above the ink tank21.5 mbar. The printed structure likewise consists of a square having an edge length of 2 cm each which has been deposited on the square with the boron ink. After the printing, the phosphorus ink is dried at temperatures between 400 C. and 600 C. on a conventional laboratory hotplate for five to ten minutes in each case. The entire structure is subsequently subjected to high-temperature diffusion in a tubular oven at 950 C. To this end, the diffusion is carried out for 30 minutes in a stream of nitrogen, followed by an oxidation process for five minutes in an atmosphere comprising nitrogen and oxygen (20% v/v) and furthermore followed by a drive-in phase of ten minutes in a nitrogen atmosphere. The diffused wafers are subsequently freed from the printed-on dopant sources by means of etching in dilute hydrofluoric acid, and the doping profile is measured in the printed areas with the aid of electrochemical capacitance-voltage measurement (ECV).

[0089] FIG. 14 shows the ECV profile of a silicon wafer which has been printed with the boron ink according to the invention, subsequently overcoated with phosphorus ink and brought to diffusion. The profile shows the boron doping arising on use of a print resolution of 508 dpi and a printing temperature of 50 C. No phosphorus doping was measured in the profile, for example change of the charge carrier type in different regions of the profile.

TABLE-US-00001 TABLE 1 Temperatures and print resolutions for the overcoating experiments described above with the boron ink according to the invention by means of a phosphorus ink likewise printed by means of ink-jet printing. Analysis of the printed and doped sandwich structures in all cases resulted in boron doping achieved in the silicon wafer. Temperature 40 50 [ C.] Boron Print resolution [dpi]: 400 450 400 Temperature Boron [ C.] Phosphorus Phosphorus 50 508 B B B 40 508 B B B 40 450 B B B 40 400 B B B 40 350 B B B 40 300 B B B 40 250 B B B 40 200 B B B

Example 5

[0090] 2 g of boron oxide are dissolved in 10.5 g of tetrahydrofuran, 3 g of acetic anhydride and 4 g of tetraethyl orthosilicate in a round-bottomed flask with stirring and refluxing and left to react for 30 minutes. 41 g of ethylene glycol monobutyl ether, in which 2 g of salicylic acid and 0.6 g of acetylacetone have been pre-dissolved, and 20 g of diethylene glycol monoethyl ether are subsequently added. When the reaction mixture has mixed completely, 10 g of aluminium tri-sec-butylate are added to this solution. The solution is refluxed for a further hour, and readily volatile solvents and reaction products are subsequently stripped off in a rotary evaporator at 60 C. with achievement of a final pressure of 50 mbar. The boron ink is subsequently filtered through a filter having a pore size of 0.45 m and deaerated. p-type test wafers which have been polished on one side are subsequently coated by means of the spin-coating process using a two-step coating programme: spinning for 15 s at 500 rpm in order to distribute the ink, followed by 2,000 rpm for 45 s. The coated wafers are subsequently dried at 500 C. on a conventional laboratory hotplate for five minutes. After this coating, the wafers are re-coated on the side already coated with boron ink with a phosphorus-containing doping ink, in accordance with the composition as mentioned in the patent application WO 2014/101990, with the aid of the same spinning programme, after which the wafers are likewise dried at 500 C. for five minutes. The double-coated wafers are brought to diffusion in a tubular oven at 930 C. in a stream of nitrogen for 30 minutes. After the diffusion, the residues of the doping media are removed from the surface with the aid of dilute hydrofluoric acid, and the wafers are tested with respect to their respective doping profiles with the aid of electrochemical capacitance-voltage measurement (ECV) and secondary ion mass spectrometry (SIMS).

[0091] The boron ink is impermeable to diffusion of phosphorus from the phosphorus ink.

[0092] FIG. 15 shows the ECV profile of a silicon wafer which has been coated with the boron ink according to the invention, subsequently overcoated with phosphorus ink and brought to diffusion. No phosphorus doping was measured in the profile, for example change of the charge carrier type in different regions of the profile. For comparison, an ECV profile of a reference sample which has been treated in the same way is depicted.

[0093] FIG. 16 shows the SIMS profile of a sample comparable to FIG. 14. The SIMS profile shows the changes in concentration of boron and phosphorus in silicon. The phosphorus profile reaches a concentration of 1*10.sup.16 atoms/cm.sup.3 after a depth of 40 nm.

[0094] In a comparative experiment, a boron ink, exclusively based on a hybrid sol consisting of precursors of silicon dioxide and boron oxide, is prepared in accordance with the procedure mentioned above. The aluminium oxide component is replaced here by silicon dioxide. The ink is applied by means of spin coating using the coating programme already mentioned and subsequently likewise overcoated with phosphorus ink. The double-coated samples are subjected to the diffusion already described and subsequently analysed in the same way.

[0095] FIG. 17 shows the ECV profile of a silicon wafer which has been coated with a boron ink which is not according to the invention, subsequently overcoated with phosphorus ink and brought to diffusion. Exclusively phosphorus doping (blue) was measured in the profile. For comparison, an ECV profile of a reference sample which has been treated in the same way is depicted.

[0096] FIG. 18 shows the SIMS profile of a sample comparable to FIG. 16. The SIMS profile shows the changes in concentration of boron and phosphorus in silicon. The concentration prevailing in the silicon was phosphorus.

[0097] The hybrid sol on the basis of silicon dioxide and boron oxide and simultaneous absence of aluminium oxide was not impermeable to diffusion of phosphorus from the phosphorus ink.

Example 6

[0098] 72.2 g of ethylene glycol monobutyl ether (EGB), 2.9 g of dimethyldimethoxysilane and 12.8 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.2 g of glacial acetic acid and 0.38 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 15.3 g. 84 g of -terpineol (isomer mixture) and 3.76 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 8.5 Pa*s at a shear rate of 25 1/s and a temperature of 23 C.

[0099] The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-polishing, using the following printing parameters:

a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber of Shore hardness of 65

[0100] The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m. The paste transfer rate is 0.93 mg/cm.sup.2.

[0101] FIG. 19 shows a photomicrograph of a line screen-printed with a doping paste according to Example 6 and dried.

[0102] FIG. 20 shows a photomicrograph of a paste area screen-printed with a doping paste according to Example 6 and dried.

[0103] FIG. 21 shows a photomicrograph of a paste area screen-printed with a doping paste according to Example 6 and dried.

Example 7

[0104] 72.3 g of ethylene glycol monobutyl ether (EGB), 2.8 g of dimethyldimethoxysilane and 12.9 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.1 g of glacial acetic acid and 1.5 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 15.3 g. 84 g of -terpineol (isomer mixture) and 3.8 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 6.7 Pa*s at a shear rate of 25 1/s and a temperature of 23 C. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m. The paste transfer rate is 1.04 mg/cm.sup.2.

[0105] FIG. 22: Photomicrograph of a line which has been screen-printed with a doping paste according to Example 7 and dried.

[0106] FIG. 23: Photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 7 and dried.

[0107] FIG. 24: Photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 7 and dried.

[0108] In a further procedure, both CZ n-type silicon wafers which have been subjected to alkaline polish-etching and also those which have been alkaline-textured and subsequently polished by means of acidic etches on one side are printed with the doping paste approximately over the entire surface (93%). The printing is carried out using a screen with stainless-steel fabric (400/18, 10 m emulsion thickness over the fabric). The paste application rate is 0.9 mg/cm.sup.2. The wafers are dried at 400 C. on a hotplate for three minutes and subsequently subjected to co-diffusion at a plateau temperature of 950 C. for 30 minutes. During the co-diffusion, the wafer is diffused and doped with boron on the side printed with the boron paste, whereas the wafer side or surface that is not printed with boron paste is diffused and doped with phosphorus. The phosphorus diffusion is in this case achieved with the aid of phosphoryl chloride vapour, which is introduced into the hot oven atmosphere transported by a stream of inert gas. As a consequence of the high temperature prevailing in the oven and the oxygen simultaneously present in the oven atmosphere, the phosphoryl chloride is combusted to give phosphorus pentoxide. The phosphorus pentoxide precipitates in combination with a silicon dioxide forming on the wafer surface owing to the oxygen present in the oven atmosphere. The mixture of the silicon dioxide with the phosphorus pentoxide is also referred to as PSG glass. The doping of the silicon wafer takes place from the PSG glass on the surface. On surface regions on which boron paste is already present, a PSG glass can only form on the surface of the boron paste. If the boron paste acts as diffusion barrier against phosphorus, phosphorus diffusion cannot take place at points at which boron paste is already present, but instead only diffusion of boron itself which diffuses out of the paste layer into the silicon wafer. This type of co-diffusion can be carried out in various embodiments. In principle, the phosphoryl chloride can be combusted in the oven at the beginning of the diffusion process. The beginning of the process in the industrial production of solar cells is generally taken to mean a temperature range between 600 C. and 800 C., in which the wafers to be diffused can be introduced into the diffusion oven. Furthermore, combustion can take place in the oven cavity during heating of the oven to the desired process temperature. Phosphoryl chloride can accordingly also be introduced into the oven during holding of the plateau temperature, and also during cooling of the oven or perhaps also after a second plateau temperature, which may be higher and/or also lower than the first plateau temperature, has been reached. Of the above-mentioned possibilities, any desired combinations of the phases of possible introduction of phosphoryl chloride into the diffusion oven can also be carried out, depending on the respective requirements. Some of these possibilities have been sketched. In this figure, the possibility of use of a second plateau temperature is not depicted.

[0109] The wafers printed with the boron paste are subjected to a co-diffusion process, as depicted in FIG. 25, in which the phosphoryl chloride is introduced into the diffusion oven before the plateau temperature which is necessary in order to achieve boron diffusion, in this case 950 C., has been reached. During the diffusion, the wafers are arranged in pairs in the process boat in such a way that their sides printed with boron paste in each case face one another (cf. FIG. 27). In each case, a wafer is accommodated in a slot of the process boat. The nominal separation between the substrates is thus about 2.5 mm. After the diffusion, the wafers are subjected to a glass etch in dilute hydrofluoric acid and their sheet resistances are subsequently measured by means of four-point measurement. The side of the wafer diffused with the boron paste has a sheet resistance of 35/ (range: 10/, whereas the opposite side of the wafer printed with the boron paste, has a sheet resistance of 70/. With the aid of a p/n tester, it is demonstrated that the side that has a sheet resistance of 35/ is exclusively p-doped, i.e. doped with boron, while the opposite side, which has a sheet resistance of 70/, is exclusively n-doped, i.e. doped with phosphorus. There is no fundamental difference between the sheet resistances on the wafers subjected to alkaline polish-etching and those in which the alkaline texture has been subjected to subsequent acidic polishing on one sideboth on the wafer side doped with phosphorus and also on the wafer side doped with boron.

[0110] In a further, identical embodiment of the co-diffusion experiment outlined above, the wafers are arranged in the process boat for diffusion in such a way that the wafer side printed with the boron paste is opposite an unprinted wafer surface (cf. FIG. 28). After the post-diffusive aftertreatment of the wafers already outlined above, the sheet resistances and also the prevailing dopings are determined using the methods already mentioned. A sheet resistance of 37/ (range: 10/) is determined on the side printed with boron paste. This side is exclusively p-doped, while a sheet resistance of 70/ is measured on the back surface. The back surface is exclusively n-doped.

[0111] In a further embodiment of the co-diffusion experiment already described, wafers are printed with the boron paste according to the invention with a paste application rate of 0.7 mg/cm.sup.2 and subjected to the same diffusion conditions. The arrangement of the wafers in the process boat was carried out in accordance with FIG. 29. After processing of the diffused wafers in accordance with the procedure already outlined, a sheet resistance of 37/ (range: 8/) can be determined on the side printed with the boron paste. The doping prevailing on this wafer surface is p-type.

[0112] In a further embodiment of the co-diffusion experiment, wafers are printed with the boron paste according to the invention with a paste application rate of 0.9 mg/cm.sup.2. The printed wafers are dried at 400 C. on a hotplate for three minutes and subsequently in a through-flow oven for a further 20 minutes. The wafers are subjected to a co-diffusion experiment already described above, where the wafer surfaces printed with boron paste are in each case arranged opposite one another.

[0113] After the diffusion, the wafers are treated further in the usual manner, and the sheet resistance on the side printed with the paste is subsequently determined by means of four-point measurement. The sheet resistance is 41/ (range: 5/). The most intensive drying of the paste results in a significant reduction in the variance of the sheet resistance.

[0114] In a further embodiment of the co-diffusion, wafers are printed with a screen using a structured screen layout. The screen used corresponds to the characteristics already mentioned above. Furthermore, the screen has a busbar to be printed centrally onto the wafer surface, from which bars or fingers with a width of 700 m each branch off both to the right and also to the left.

[0115] Lands with a width of 300 m, which protect the wafer surface against the paste print, are located between the bars. The wafers printed in this way are dried at 400 C. on a hotplate for three minutes, subsequently aligned in the process boat in an arrangement in accordance with FIG. 28 and brought to diffusion using the co-diffusion conditions already described. After the further treatment of the diffused wafers, which has already been described above, these are investigated by means of an imaging process in order to determine the sheet resistance (sheet resistance imaging). The measurement method used was carried out without calibration, which means that the sheet resistance information is coded via the signals and signal changes obtained by the measurement method (in this case count rates of the IR radiation detected with the aid of the process). FIG. 29 shows a photomicrograph of a printed structure after the co-diffusion. The boron paste is still present on the sample. The regions printed with the boron paste have a dark-blue colour. A wafer which has been printed with the boron paste according to the invention and dried is depicted: the structure shown in this figure corresponds in principle to the structuring described above, apart from the fact that the busbar arranged centrally is not present on this wafer surface and the dimensions of the printed bars do not correspond exactly to the dimensions likewise already mentioned. The structures evident from FIG. 29 are transferred to the sheet resistance mapping shown in FIG. 30. The regions depicted in orange-red in FIG. 30 correspond to the structures printed with the boron paste, while the red regions can be assigned to the recesses in the structure which have been diffused with phosphorus as a consequence of the co-diffusion. In addition, it is evident that the adjacent regions have a very sharp delimitation and a very well-defined transition region, with the measurement accuracy of the means available and taking into account the measured wafer surfaces (scattering of the signal). FIG. 30 shows sharply delimited p/n junctions, alternating in accordance with the performance of the experiments, which can be produced with the aid of a single high-temperature diffusion step.

[0116] Furthermore, an ECV profile (electrochemical capacitance/voltage profiling) of the busbar region produced with the aid of the boron paste according to the invention and shown in FIG. 29 and also in FIG. 30 is depicted in FIG. 31. The profile shows an emitter profile having a depth of the p/n junction of about 600 nm. Exclusively p-type doping prevails. The surface concentration of the charge carriers (holes) of the emitter profile is about 2*10.sup.20 cm.sup.3. Through suitable manipulation of the process procedure of the co-diffusion process, both the depth of the profile and also the surface concentration can be adjusted in the desired manner (for example through the choice of the diffusion temperature, the diffusion length, the composition of the gas atmosphere used during the diffusion process, and here in particular by setting a determined oxygen concentration).

[0117] FIG. 25 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 7 and dried.

[0118] FIG. 26 shows the photograph of a monocrystalline silicon wafer printed with boron paste according to Example 7 in the form of a bar structure.

[0119] FIG. 27 shows an arrangement of wafers in a process boat during a co-diffusion process. The wafer surfaces printed with boron paste are opposite one another.

[0120] FIG. 28 shows an arrangement of wafers in a process boat during a co-diffusion process. The wafer surfaces printed with boron paste are opposite one another.

[0121] FIG. 29 shows a photomicrograph of a wafer printed with the boron paste according to the invention (alkaline-textured wafer surface subjected to acidic post-polishing on one side). The nominal dimensions of the printed-on structure are mentioned in the text.

[0122] FIG. 30 shows a sheet resistance mapping with the aid of the SRI process. The orange-yellow regions correspond to boron doping, whereas the red regions can be assigned to phosphorus doping. The structure shown corresponds to that depicted in FIG. 29.

[0123] FIG. 31 shows an ECV profile of an emitter profile (boron, p-type) obtained by means of the boron paste according to the invention and using a co-diffusion process. The depth of the p/n junction is about 600 nm. The surface concentration of the charge carriers (holes) is about 2*10.sup.20 cm.sup.3.

Example 8

[0124] 563.2 g of ethylene glycol monobutyl ether (EGB), 23 g of dimethyldimethoxysilane and 102.2 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 97.8 g of glacial acetic acid and 6 g of acetaldoxime are added to this mixture in the said sequence with stirring. 15.6 g of water, dissolved in 50 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 119.6 g. 782 g of -terpineol (isomer mixture) and 32.2 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 7.5 Pa*s at a shear rate of 25 1/s and a temperature of 23 C.

[0125] The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters:

a screen separation of 2 mm,
a printing speed of 200 mm/s,
a flooding speed of likewise 200 mm/s,
a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65.

[0126] The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m.

[0127] The paste transfer rate is 1.17 mg/cm.sup.2.

[0128] FIG. 32 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 8 and dried.

[0129] FIG. 33 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 8 and dried.

[0130] FIG. 34 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 8 and dried.

Example 9

[0131] 72.5 g of ethylene glycol monobutyl ether (EGB), 2.9 g of dimethyldimethoxysilane and 12.8 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.2 g of glacial acetic acid and 1.4 g of 1,3-cyclohexanedione are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 20.8 g. 103 g of -terpineol (isomer mixture) and 3.9 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 19.3 Pa*s at a shear rate of 25 1/s and a temperature of 23 C. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m. The paste transfer rate is 1.15 mg/cm.sup.2.

[0132] FIG. 35 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 9 and dried.

[0133] FIG. 36 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 9 and dried.

[0134] FIG. 37 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 9 and dried.

Example 10

[0135] 74.2 g of ethylene glycol monobutyl ether (EGB), 2.9 g of dimethyldimethoxysilane and 12.9 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.3 g of glacial acetic acid and 1.9 g of 3,5-dihydroxybenzoic acid are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 20.5 g. 96.5 g of -terpineol (isomer mixture) and 3.9 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 23.8 Pa*s at a shear rate of 25 1/s and a temperature of 23 C. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m. The paste transfer rate is 1.08 mg/cm.sup.2.

[0136] FIG. 38 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 10 and dried.

[0137] FIG. 39 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 10 and dried.

[0138] FIG. 40 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 10 and dried.

Example 11

[0139] 72.2 g of ethylene glycol monobutyl ether (EGB), 2.9 g of dimethyldimethoxysilane and 12.8 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.3 g of glacial acetic acid and 1.6 g of salicylic acid are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 16.9 g. 94.5 g of -terpineol (isomer mixture) and 4 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 8.4 Pa*s at a shear rate of 25 1/s and a temperature of 23 C. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m. The paste transfer rate is 0.92 mg/cm.sup.2.

[0140] FIG. 41 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 11 and dried.

[0141] FIG. 42 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 11 and dried.

[0142] FIG. 43 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 11 and dried.

Example 12

[0143] 72.3 g of ethylene glycol monobutyl ether (EGB), 2.9 g of dimethyldimethoxysilane and 12.9 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.2 g of glacial acetic acid and 0.56 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 21.6 g. 108 g of -terpineol (isomer mixture) and 3.9 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 9.3 Pa*s at a shear rate of 25 1/s and a temperature of 23 C. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m. The paste transfer rate is 1.06 mg/cm.sup.2.

[0144] FIG. 44 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 12 and dried.

[0145] FIG. 45 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 12 and dried.

[0146] FIG. 46 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 12 and dried.

Example 13

[0147] 72.1 g of ethylene glycol monobutyl ether (EGB), 2.9 g of dimethyldimethoxysilane and 12.8 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.3 g of glacial acetic acid and 1.1 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 19.3 g. 99 g of -terpineol (isomer mixture) and 3.9 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 5.9 Pa*s at a shear rate of 25 1/s and a temperature of 23 C. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m. The paste transfer rate is 0.86 mg/cm.sup.2.

[0148] FIG. 47 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 13 and dried.

[0149] FIG. 48 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 13 and dried.

[0150] FIG. 49 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 13 and dried.

Example 14

[0151] 577.2 g of ethylene glycol monobutyl ether (EGB), 19.9 g of tetraethyl orthosilicate, 19.9 g of 1,2-bis(triethoxysilyl)ethane and 102.2 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 97.8 g of glacial acetic acid and 6 g of acetaldoxime are added to this mixture in the said sequence with stirring. 15.6 g of water, dissolved in 50 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 269.7 g. 782 g of -terpineol (isomer mixture) and 32.2 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively, and finally a further 100 ml of ethylene glycol monobutyl ether are added. A pseudoplastic and very readily printable, transparent paste forms. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m.

[0152] FIG. 50 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 14 and dried.

Example 15

[0153] 60 g of ethylene glycol monobutyl ether (EGB), 1.7 g of tetraethyl orthosilicate, 1.4 g of 1,2-bis(triethoxysilyl)ethane, 1 g of dimethyldimethoxysilane and 12.8 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.4 g of glacial acetic acid and 0.7 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 17.5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 15.2 g. 85 g of -terpineol (isomer mixture) and 4 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively, and finally a further 100 ml of ethylene glycol monobutyl ether are added. A pseudoplastic and very readily printable, transparent paste forms. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m.

[0154] FIG. 51 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 15 and dried.

[0155] FIG. 52 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 15 and dried.

Example 16

[0156] 72 g of ethylene glycol monobutyl ether (EGB), 1.3 g of 1,2-bis(triethoxysilyl)ethane, 2 g of dimethyldimethoxysilane and 12.8 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.2 g of glacial acetic acid and 0.7 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 15.3 g. 92 g of -terpineol (isomer mixture) and 3.9 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively, and finally a further 100 ml of ethylene glycol monobutyl ether are added. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 10.6 Pa*s at a shear rate of 25 1/s and a temperature of 23 C. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m.

[0157] FIG. 53 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 16 and dried.

[0158] FIG. 54 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 16 and dried.

[0159] FIG. 55 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 16 and dried.

Example 17

[0160] 72.2 g of ethylene glycol monobutyl ether (EGB), 0.4 g of 1,2-bis(triethoxysilyl)ethane, 2.6 g of dimethyldimethoxysilane and 12.9 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.4 g of glacial acetic acid and 0.7 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. The mass loss of readily volatile reaction products is 15.3 g. 92 g of -terpineol (isomer mixture) and 4 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively, and finally a further 100 ml of ethylene glycol monobutyl ether are added. A pseudoplastic and very readily printable, transparent paste forms. The viscosity of the paste is 5.4 Pa*s at a shear rate of 25 1/s and a temperature of 23 C. The paste is printed with the aid of a screen printer using a trampoline screen with stainless-steel fabric (400 mesh, 18 m wire diameter, calendered, 8-12 m emulsion on top of the fabric) onto wafers which have been subjected to alkaline polish-etching, using the following printing parameters: a screen separation of 2 mm, a printing speed of 200 mm/s, a flooding speed of likewise 200 mm/s, a doctor-blade pressure of 60 N during the printing operation and a doctor-blade pressure of 20 N during the flooding, and using a carbon fibre doctor blade with polyurethane rubber having a Shore hardness of 65. The printed wafers are subsequently dried in a through-flow oven warmed to 400 C. The belt speed is 90 cm/s. The length of the heating zones is 3 m.

[0161] FIG. 56 shows a photomicrograph of a line which has been screen-printed with a doping paste according to Example 17 and dried.

[0162] FIG. 57 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 17 and dried.

[0163] FIG. 58 shows a photomicrograph of a paste area which has been screen-printed with a doping paste according to Example 17 and dried.

Example 18

[0164] 72.2 g of ethylene glycol monobutyl ether (EGB), 3.55 g of trimethoxyvinylsilane and 12.9 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.4 g of glacial acetic acid and 0.7 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. 97 g of -terpineol (isomer mixture) and 4 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms.

Example 19

[0165] 72.2 g of ethylene glycol monobutyl ether (EGB), 5.85 g of dimethoxydiphenylsilane and 12.9 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.4 g of glacial acetic acid and 0.7 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. 97 g of -terpineol (isomer mixture) and 4 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms.

Example 20

[0166] 72.2 g of ethylene glycol monobutyl ether (EGB), 3.53 g of bis(dimethoxy-dimethylsilyl)-1,2-ethane and 12.9 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.4 g of glacial acetic acid and 0.7 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. 97 g of -terpineol (isomer mixture) and 4 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms.

Example 21

[0167] 72.2 g of ethylene glycol monobutyl ether (EGB), 4.36 g of dimethoxymethyl-phenylsilane and 12.9 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.4 g of glacial acetic acid and 0.7 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. 97 g of -terpineol (isomer mixture) and 4 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms.

Example 22

[0168] 72.2 g of ethylene glycol monobutyl ether (EGB), 5.8 g of triethoxyphenylsilane and 12.9 g of aluminium tri-sec-butylate (ASB) are initially introduced in a glass flask and stirred until a homogeneous mixture forms. 12.4 g of glacial acetic acid and 0.7 g of acetaldoxime are added to this mixture in the said sequence with stirring. 1.9 g of water, dissolved in 5 g of EGB, are subsequently added dropwise, and the mixture is refluxed at 80 C. for four hours. After the warming, the hybrid sol is subjected to vacuum distillation at 70 C. for two hours until a final pressure of 20 mbar has been reached. 97 g of -terpineol (isomer mixture) and 4 g of tetraacetoxy diborate are subsequently added to the distilled mixture and mixed intensively. A pseudoplastic and very readily printable, transparent paste forms.