T 0942/06 (Low-e glass/GUARDIAN INDUSTRIES CORP.) 21-09-2011

Reasons for the Decision

1. Amendments The opposing parties filed numerous objections under Article 123(2) EPC, both against the claims as granted and against the amended claims in accordance with auxiliary requests 1 to 4. The board found none of these objections well-founded and decided in the oral proceedings that the requirements of Article 123(2) EPC were met. It is not necessary to provide a detailed reasoning in this respect here because the patent cannot be maintained for other reasons.

2. Admissibility of certain submissions Experimental report V1 to V4 concerns a reworking of examples 1 and 2 of the opposed patent, submitted by appellant 1 with the statement of grounds of appeal. The board considers that these new submissions were made in an effort to invalidate the arguments of the opposition division in the contested decision. They are thus admissible.

3. Sufficiency of disclosure (Article 83 EPC)

3.1 The requirement of sufficiency of disclosure According to the established jurisprudence of the boards of appeal, the requirement of sufficiency of disclosure is only met provided the invention as defined in the independent claim can be performed by the person skilled in the art within the whole area claimed without the burden of an undue amount of experimentation, taking into consideration common general knowledge and the whole information content of the patent in suit (see decision T 435/91, OJ EPO 1995, 188, point 2.2.1, third paragraph, of the Reasons; and T 409/91, OJ EPO 1994, 653, point 2, first paragraph, penultimate sentence).

3.2 The invention

3.2.1 The claimed invention aims at providing coating systems for glass substrates which exhibit high visible transmission, very low emissivity values and which are substantially neutral in colour (see paragraphs [0001] and [0017]). Additionally, claim 1 as granted also recites as desiderata that the visible transmittance (Tvis) of the coating system, when applied to a clear glass having a thickness comprised between 2 to 6 mm, is at least 84%, the sheet resistance (Rs) is less than or equal to 5.5 ohms/sq. and the normal emissivity (En) is less than or equal to 0.065.

The layer system in accordance with the claimed invention thus aims at achieving a combination of high Tvis and at the same time low En and low Rs (see paragraph [0037]). Moreover, although not specifically claimed, the sputter coated glass article should have a substantially neutral visible reflected colour (i.e. colourless to slightly blue and non-purple) when viewed from the glass side, and a substantially non-mirror-like reflectance (see paragraphs [0003] and [0040]). The skilled person thus has to arrive at the simultaneous achievement of these properties.

3.2.2 In order to obtain these objectives, a multi-layer sputter-coated system according to claim 1 comprises the following essential layers a), c), d) and e) from the glass outwardly, defined in structural terms: a) an undercoat layer formed of at least a metal oxide or nitride having an index of refraction comprised in the interval 2.35 - 2.75 and a thickness of 16 to 32 nm; c) a layer of metallic silver; d) an upper intermediate layer located between said silver layer and said overcoat layer, wherein said upper intermediate layer is selected from an oxide or a nitride of Al, Ti, Zn, Zr, Cr, Ta, Mg or mixtures thereof; and e ) an overcoat layer formed of at least a metal oxide or nitride having an index of refraction comprised in the interval 1.85 - 2.25. Claim 1 also comprises, as functional features or desiderata, the requirements that: f) said layer of metallic silver and said overcoat layer are chosen in combination of a thickness sufficient such that said glass article, when said glass substrate consists of a clear glass of a thickness comprised between 2-6 mm, has a visible transmittance (Tvis) of at least 84%, a sheet resistance (Rs) of less than or equal to 5.5 ohms/sq., and a normal emissivity (EN) of less than or equal to 0.065. Therefore, in accordance with claim 1 as granted, the layers and their respective thicknesses have to be selected - within the limits of structural claim features a) to e) - such that the optical and physical properties of the so coated glass article satisfy functional claim feature f).

3.2.3 It is known in the art that at least some of the above defined desired optical and physical characteristics are in mutual conflict. Therefore, a trade-off is often required for their simultaneous achievement (see paragraphs [0005] and [0006] of the opposed patent). Under these circumstances, and in view of the functional definition of the desiderata, the board considers it essential for the patent to contain sufficient and enabling information allowing the skilled person to carry out the invention and achieving the desired set of product characteristics without an undue burden of trial and error.

3.3 The gaps of information

3.3.1 In order to put the invention into practice, the appellants have identified the following alleged gaps of information:

(a) the chemical composition of the undercoat layer;

(b) the thickness of the silver layer;

(c) the thickness of the intermediate layer;

(d) the kind of metal oxide or nitride used for the overcoat layer;

(e) the thickness of the overcoat layer;

(f) the process parameters for sputtering an overcoat layer of a metal oxide or nitride having an index of refraction comprised in the interval of 1.85 - 2.25.

3.3.2 The appellants also criticized that the working examples of the patent were not reproducible and thus did not provide a suitable starting point for the skilled person trying to exploit the invention over the full width of the claims.

3.4 Guidance in the description As to the various parameters enumerated in point 3.3.1 above, the description of the opposed patent provides in particular the following information: - A preferred material for the undercoat layer is TiO2. The thickness of the layer is 16 to 32 nm, preferably 20 to 30 nm [paragraph 0043]); - The thickness of the silver layer is preferably 10 to 18 nm, more preferably 14 to 17 nm (paragraph [0042]); - A preferred material for the overcoat layer is SnO2. The thickness of the layer is 35 to 70 nm, preferably 40 to 55 nm [paragraph 0044]); - The overcoat layer is preferably made of TiO2, having a thickness of preferably 0.5 to 1.5 nm, most preferably of 1.0 nm (paragraph [0046]); - Typical process parameters (target composition, reactive and inert gas flows, pressure, cathode power, cathode voltage and current, line speed and number of passes) for sputtering the individual layers are indicated in connection with examples 1 and 2 (see Tables 1 and 2; Figures 1 and 2). According to paragraph [0048], the process and apparatus used to form the various layers are conventional and employ operating parameters well known to those of skill in the art.

3.5 Instruction by way of examples

3.5.1 The opposed patent contains also two working examples to illustrate the invention (see paragraphs [0049] to [0055], [0058] and [0059]). Appellant 1 had argued already during the opposition procedure that the two working examples of the opposed patent represented in Figures 1 and 2 and described in paragraphs [0049] to [0055] could not be successfully repeated. In the appellant's view, it was in particular not possible to produce a sputter-coated glass article as per example 2 having a luminous transmittance of 89.11 % and a silver layer thickness as high as 16.5 nm. The appellant referred in this context to E7 (Figure 7) disclosing a maximum in luminous transmittance of only 83% at a silver layer thickness of 12 nm. Appellant 2 concurred with these arguments. However, in the contested decision (Reasons, point 21) the opposition division dismissed the appellant's objection on the grounds that the explanations based on E7 were inconclusive and that a faithful reworking of the examples had not been attempted.

3.5.2 Examples 1 and 2 The examples are discussed in more detail in the following (cf. also point 3.5.5., Table 1). Example 1 (see paragraph [0049]): soda-lime silica float glass (2 mm) 29 nm TiO2, refraction index n = 2.56 (undercoat) metallic Ag (16.5 nm) 1 nm TiO2 (protective intermediate layer) 48 nm SnO2 (overcoat) The sputtering conditions for the various layers of example 1 are reported in Table 2 (page 9 of the patent in suit). Properties of the coated glass (see paragraph [0055]): Tvis: 87.15 % En: 0.061 Rs: 4.87 ohm/sq Color: substantially neutral

Example 2 (see paragraph [0050]): soda-lime silica float glass (3.1 mm) 22.4 nm TiO2, refraction index n = 2.56 (undercoat) 5.2 nm ZnO (barrier layer) metallic Ag (16.5 nm) 1 nm TiO2 (protective intermediate layer) 48 nm SnO2 (overcoat) The sputtering conditions for the various layers of example 2 are reported in Table 1 (page 7 of the patent in suit). Properties of the coated glass (see paragraph [0053]): Tvis: 89.11 % En: 0.054 Rs: 4.80 ohm/sq Color: substantially neutral These experimental results, in particular the values for Tvis, En and Rs, fall within the scope of the independent claims.

3.5.3 Reproducibility According to appellant 1, the examples 1 and 2 did not provide the skilled person with a reliable starting point for determining the necessary process parameters, because they could not be reproduced successfully. The appellant also strictly denied the possibility of obtaining - by standard sputtering techniques - a low-absorbing TiO2 layer having a refractive index of as high as 2.56 at a wavelength of 550 nm. It was asserted that any TiO2 layers obtainable by sputtering inevitably had an increased light absorption and an increased extinction coefficient k, contrary to the requirement in the opposed patent according to which k should be less than 0.01 (see paragraph [0045]).

3.5.4 Re-working of the examples In support of its assertions, appellant 1 filed a report comprising the sputter-coated glass samples V1 to V4, reproducing examples 1 and 2 of the opposed patent (see appeal brief, pages 9 to 14). These test samples V1 to V4 were prepared on a laboratory-scale sputter-coating apparatus using a glass substrate ("Optifloat") of 2.1 mm thickness having a light transmittance of TL = approx. 90%, i.e. higher than of the glass used in example 2 of the opposed patent. According to experiments V1 and V2, the TiO2 layers were prepared by reactive sputtering from Ti targets, yielding TiO2 having a refractive index of approximately only 2.45 (550 nm) and k = 0.001 (550 nm) (i.e. outside claim 1). In accordance with experiments V3 and V4, the TiO2 layers were prepared by non-reactive sputtering from Ti targets to obtain metallic Ti layers and subsequently oxidizing the metal at 500 ºC in air, yielding TiO2 layers having a refractive index of approximately 2.58 (550 nm) and k = 0.001 (550 nm). The optical properties of these TiO2 layers thus corresponded to those of the opposed patent and also to the requirement of claim 1. The layer thicknesses and the sputter conditions (except for TiO2 in examples V3 and V4; see above) - save for minor experimental deviations - sufficiently matched those of examples 1 and 2 of the opposed patent (see the appeal brief of appellant 1, page 12, Table 3; for instance in V3: first TiO2 layer having a thickness of 20.8 nm was produced compared with a nominal thickness of 22.4 nm). The board is satisfied that the deviations are within normal experimental margins of variation and sufficiently small to allow a meaningful comparison with the examples of the opposed patent.

3.5.5 Comparative results The measured values of the optical properties of samples V1 to V4, in comparison with examples 1 and 2 of the opposed patent, are presented in the following Table 1: Table 1

| TVIS %| RG %| RF %|RS Omega | EN|

| | | | |/ ? | |

| | | | | | |

| V1 | 78.4 | 15.9 | 13.1 | 2.4 | 0.031 |

| | | | | | |

| | | | | | |

| V3 | 81.1 | 12.1 | 10.0 | 2.8 | 0.033 |

| | | | | | |

| | | | | | |

|Example 1| 89.11 | 5.3 | 4.28 | 4.8 | 0.054 |

| | | | | | |

| | | | | | |

| V2 | 80.0 | 12.8 | 10.1 | 3.4 | 0.039 |

| | | | | | |

| | | | | | |

| V4 | 81.4 | 9.5 | 7.2 | 3.4 | 0.036 |

| | | | | | |

| | | | | | |

|Example 2| 87.15 | 6.19 | 4.59 | 4.87 | 0.061 |

| | | | | | |

| | | | | | |

3.5.6 Comments on the comparative results a) Visible light transmittance Tvis It is evident that the transmissivity values Tvis of samples V1 to V4 are significantly inferior to those of the examples of the opposed patent. Similar significant discrepancies exist as regards the values for RS and EN. These differences are by far greater than one would expect from the small variances in the nominal thicknesses of the coating layers. More importantly, the transmissivity values Tvis found in all examples V1 to V4 fall substantially below the claimed minimum value of 84%, irrespective of whether the refractive index n (550 nm) of the TiO2 layer was low (n= 2.45 as in V1, V2) or high (n= 2.58 as in V3, V4). This suggested that the claimed invention could not be reproduced, no matter what the refractive index of the TiO2 layers was. The data demonstrated that the attempt of reproducing the examples of the opposed patent failed to achieve the high visible transparency values of at least 84% as set out in the claims, even though the glass substrate used by appellant 1 had been more light transmissive than the float glass used in the examples of the patent. Appellant 1 asserted that it was neither plausible per se nor derivable from the patent that coatings having a Ag layer thickness at the higher end of the patent's preferred range (10 to 18 nm) could yield the desired high visible transmittance. Rather, the literature at hand and the experimental evidence suggested that this was only possible with a Ag layer thickness of around 12 nm. b) Colour Appellant 1 and appellant 3 both argued that the neutral reflexion colour (i.e. colourless to slightly blue and non-purple) when viewed from the glass side, had also not been achieved by the examples of the opposed patent, as could be seen from the respective bH-values of -8.79 and -8.44 (colour coordinates, III.C. Hunter).

3.6 Conclusions of the appellants In summary, according to the appellants, the patent did not provide a solution to the underlying problem over the entire claimed range of thicknesses of 16 to 32 nm for the underlayer, but only over a very limited range of thicknesses, in combination with undisclosed limited thickness ranges for the silver layer. Even these limited combinations required special undisclosed selections of the material of the overcoat and of its thickness, in order to simultaneously achieve a high transmittance, a low emissivity and a neutral colour tone. Appellants 1 and 3 therefore submitted that the requirement of Article 83 EPC was not met.

3.7 The respondent's position and the board's view a) Visible light transmittance Tvis The respondent argued that the failure of reproducing the high visible transmittance was possibly attributable to a high content of metallic Ti in the upper intermediate TiO2 layer, causing high absorption. Partial or full oxidation of the TiOx layer sputtered in pure Ar atmosphere occurred only when the next layer was sputtered in a highly oxidizing atmosphere. However, the extent of oxidation of the TiOx layer in appellant 1's examples was not known and might have been incomplete. However, the board notes that it was undisputed that a partial oxidation of the TiOx layer could be expected only if the thickness of the barrier layer was substantially greater than 1 nm. An incomplete oxidation can also be effectively ruled out in view of the observed transmission Tvis, absorption AG and reflection RF values (which add up to 100%). In examples V1 to V4, one observes AG values higher than those of the opposed patent by only 0.1% to 3%, which differences are too small to explain the differences in Tvis of 6% to 11%. The board therefore accepts the assertion of appellant 1 that the TiO2 layers of V1 to V4 were substantially fully oxidised, as required by the opposed patent. b) Silver layer thickness According to the respondent, it had not been disputed that the claimed invention worked with a thinner Ag layer of 12 nm. Under such circumstances, it was not plausible that the patentee should have included two non-working examples in the description which, due to their relatively high Ag layer thickness, would exhibit the desired high visible transmittance values. The board cannot accept this argument because from the fact that the claimed invention worked with a thinner Ag layer of 12 nm no conclusion can be drawn for higher Ag layer thicknesses. It is undisputed that thicker Ag layers make it increasingly difficult to maintain high visible transmittance. Therefore, the experimental failure of V1 to V4 is not at variance with a positive result at 12 nm. However, the patent's preferred range for the silver layer thickness is from 14 to 17 nm (see paragraph [0042]) and 16.5 nm in the examples; the patent in suit contains no working examples with a 12 nm Ag layer. The skilled person is therefore, in a first orientation, not directed by the patent at 12 nm, but at substantially higher silver layer thicknesses and thus bound to fail in the attempt to achieve the desired properties as defined in claim 1. Another argument of the respondent was that the skilled person, having failed to achieve the desired high transmissivity of at least 84% following the examples 1 or 2, would have immediately realized that the thickness of the Ag layer should be reduced in order to overcome the problem. The skilled person would thus have been directed to embodiments having a substantially thinner Ag layer and higher transmission. The board disagrees because the thickness of the Ag layer is known to have a direct influence on the sheet resistance Rs and on the emissivity EN. The patent itself states in paragraphs [0012] and [0013] that conventional coating systems having a thinner Ag layer in the range of from 5 to 15 nm may exhibit the desired high visible light transmittance (e.g. 85% to 86%) but suffer from high sheet resistance (e.g. about 6.7 to 8.2 ohms/sq) and a low level of IR reflection. Besides, such systems tend to exhibit a light purple coloration. Therefore, the skilled person would not have considered - in the board's view - thinner Ag layers as a viable alternative. c) Computer simulation of examples 1 and 2 Appellant 3 had filed under cover of a letter dated 6 October 2005 a document designated as "Annexe 4" and containing data obtained by a computer simulation of examples 1 and 2 of the opposed patent concerning the layer system glass/TiO2/Ag/TiO2/SnO2 wherein the Ag layer thickness was 16.5 nm. Said simulation gave transmissivity values Tvis of 89.11% and 87.15%, respectively, in close agreement with the values given in the opposed patent and well above the 84% minimum value. In view of these positive results, the respondent argued that re-working the patent's examples 1 and 2 by appellant 1 (V1 to V4) was flawed and not conclusive. However, as appellant 3 pointed out, the data of Annex 4 were not obtained experimentally, but as a result of a computer simulation of the claimed coating layer system. In the simulation programme, inter alia a refractive index of TiO2 of 2.56 was used as input data, in accordance with the values given in examples 1 and 2 of the opposed patent. It was, however, disputed by appellant 3 and appellant 1 that TiO2 layers having a refractive index n of 2.56 (550 nm) could be obtained by conventional sputtering (see points 3.6.8 d) and e) below). Therefore, the computer simulation results do not - in the board's view - conclusively prove that the claimed invention can be worked. In any event, the board considers that the experimental evidence, as submitted by appellant 1, carries a considerably greater persuasive weight. d) Refractive index of the TiO2 layer The respondent argued that experiments V1 and V2 had not been carried out according to the invention by appellant 1 because the refractive index of the TiO2 layer was lower than required and because experiments V3 and V4 comprised a heat treatment for obtaining the TiO2 layer. These arguments are, although factually correct, in the board's opinion besides the point. Experiments V1 and V2 demonstrate that TiO2 films having a refractive index of 2.56 (550 nm) could not be obtained by conventional sputtering, a fact which is also supported by E8 and E10 (see point 3.7 e), last paragraph). In an effort to nevertheless obtain such highly refractive TiO2 films, and to produce a coated glass article having the structural features as claimed in claim 1 of the patent, an oxidizing heat treatment was carried out in experiments V3 and V4. Therefore, the deviations from the patent's teaching were only made in a fair attempt of reproducing the examples. e) Sputtering Contrary to appellant 1's assertion, documents E5, E6 and E8 proved that TiO2 layers having a refractive index n between 2.2 and 2.8 could be obtained by RF magnetron sputtering. In the board's view, these arguments were plausibly refuted by the appellants on the ground that E5, E6 and E8 do not relate to conventional sputtering, but to high rate sputtering processes requiring special targets. According to the patent in suit (paragraph [0048]), however, the process and apparatus used to form the various layers on glass substrate is a conventional sputter-coating system using operating parameters well known to those of skill in the art. E6 reports a number of refractive indices for TiO2 in the range of from 2.488 to 2.9467, depending on the crystal modification, but does not concern layers of TiO2, but presumably bulk material. The document is moreover completely silent about the production of TiO2 by sputtering. E8 is about the preparation of TiO2 film by RF magnetron sputtering. It is reported that the crystal structure of the films and hence the refractive index strongly depended on the substrate position and sputtering gas pressure. A refractive index of 2.44 to 2.67 was measured for 100% rutile films, whereas anatase films obtained under different sputtering conditions exhibited a lower index of from 2.25 to 2.47 (page 4955, point 4, "Conclusion"; E8, page 4952, left hand column, point 3.2; page 4952, Figure 5). The highly refractive rutile modification was obtained only under specific sputtering conditions. The opposed patent is however silent about such specific sputtering conditions. E10 confirms that conventional magnetron sputtering of TiO2 layers yields only a product having a lower refractive index of 2.4 to 2.45 (550 nm). A proprietary TwinMag® process using a mid-frequency (MF) double magnetron cathode is proposed in order to obtain TiO2 films having a refractive index of 2.5 to 2.7 (550 nm). These TiO2 films show higher light transmission, colour neutrality and a compact, smoother structure (see E10, page 203, "Abstract"; page 205, left hand column, points 2.4 and 2.5). Appellant 2 had correctly pointed out that it was known from E22 (published 1997) that sputtering of highly refractive TiO2 had become commercially viable by the TwinMag® process of E10 and that it offered considerable advantages as a coating material in low-E glasses. These special deposition techniques are, however, neither explicitly nor implicitly disclosed in the opposed patent.

3.8 Lack of guidance According to the respondent the invention was sufficiently disclosed in view of the clear guidance given in the description, paragraphs [0039] and seq., even if the examples could not be repeated or only with difficulties (which was denied). The question to be answered by the board is whether there are gaps of information and whether they may be filled by the general common knowledge of the skilled person, without undue burden of trial and error. To this end, the board has to investigate whether the examples can be reproduced and whether they illustrate the invention and its desired effects, as claimed. In view of the multitude of free parameters and the functional limitations characterizing the claimed invention (see point 3.2.2, item f)), without a proper starting point, preferably in the form of working examples, the board considers the guidance in the description to be insufficient. It is from such examples illustrating the preferred embodiments of the invention that the skilled person generally sets out to explore the invention in its full ambit, as defined in the claims. However, in the present case, the examples are deficient in that they do not yield the desired results in terms of the functional claim feature f) defined under point 3.2.2. Moreover, they are misleading in that they suggest that the invention can be carried out best with layer systems having a relatively high Ag layer thickness of 16.5 nm, whereas in fact such layer systems do not solve the problem posed. In addition, the description is deficient and misleading in that the necessary highly refractive TiO2 layers cannot, in fact, be obtained by standard sputtering methods. The tests V1 to V4 carried out by appellant 1 strongly suggest that even layer systems with TiO2 layer having a high refractive index of approximately 2.58 (550 nm) do not generally (that is, in combination with other layers as per examples 1 and 2 of the opposed patent) solve the problem posed. The common general knowledge of the person skilled in the art does not help to fill the gaps of information and to cure the lack of guidance.

3.9 The board's conclusion The board concludes that the subject-matter of claim 1 of the main request is not disclosed in a manner sufficiently clear and complete for it to be carried out by a person skilled in the art. No different or further arguments were brought forward by the respondent concerning auxiliary requests 1 to 4. In fact, claim 1 of all these requests contains the same functional features in terms of Tvis, Rs and En as claim 1 of the main request, so that the same problem of simultaneously achieving these desiderata arises. Claim 1 of the third auxiliary request in addition calls for a substantially neutral, non-purple colour and a non-mirror-like reflectance of the coated glass article, thus placing further constraints on the choice of the parameters. Therefore, the arguments given above concerning claim 1 of the main request apply mutatis mutandis to the subject-matter of the claims of auxiliary requests 1 to 4. The requirements of Article 83 EPC are therefore not met.

4. As no allowable request is on file, the opposed patent must be revoked.

5. Requests for remittal (Article 111(1) EPC) and apportionment of costs (Article 104 EPC) The requests for a remittal and apportionment of costs (see the respondent's letter dated 26 April 2007, page 2, point 5) cannot be allowed. A remittal of the case to the department of first instance would have doubtlessly caused a further, substantial delay in delivering a final verdict. In view of the considerable amount of time the granting and opposition procedure have already taken (the priority date of the opposed patent goes back to the year 1997), a further delay is not in the interest of the public. An apportionment of costs was requested only in connection with further oral proceedings before the first instance. As the case is not remitted, the point of apportioning such costs is moot.