T 1516/18 (Melt composition/Rockwool) 21-07-2022

Main request

1. Novelty in view of C5

1.1 Novelty was contested in view of composition "N" of C5 as well as the general disclosure of that document.

1.2 However, the subject-matter of claim 1 differs from composition "N" in that the iron oxide content is lower, and in that the proportion of Fe(2+) based on total Fe is greater than 80%.

1.3 As regards the iron oxide content, there is no basis to view composition "N" as a target composition to which the tolerances allowed in conjunction with EUCEB certification had to be applied. There is no mention of EUCEB certification and the applicable tolerances in C5. Nor is there any reason to doubt that the Fe2O3 content of composition "N", namely 9.8 weight% expressed as Fe2O3 (calculated by appellant 3), can be distinguished from the upper limit of 9 weight% specified in claim 1 at issue; this is not a question of rounding the figure because a value of 9.8 in any case is above values that could be rounded to "9".

1.4 Furthermore, Article 69 EPC does not provide any basis for interpreting the claim such that it encompasses allegedly equivalent iron oxide contents for the purpose of assessing novelty. Appellant 3 considers the prior-art value to be equivalent to the claimed range. Irrespective of there being no evidence for this, equivalents are not taken into consideration when assessing novelty (Case Law of the Boards of Appeal of the EPO, 9th edition, 2019, I.C.4.5). This concerns not only the disclosure of the prior art but also the interpretation of the claim, as is clear from T 279/12 (Reasons 1.3) and T 2614/17 (Reasons 3.3).

1.5 C5 is silent as regards the proportion of Fe(2+). The appellants derive the presence of Fe(2+) from a cupola furnace being used. However, a cupola furnace is merely one of several alternatives in C5 (page 8, lines 10-12). There is no disclosure of composition "N" being melted using specifically a cupola furnace; in the examples, a crucible furnace is used. There is no evidence that the resulting melt exhibits the required proportion of Fe(2+).

1.6 Considering the general disclosure of C5 (see the paragraph bridging pages 3-4 and the claims), several selections are necessary to arrive at a composition within the scope of claim 1, in particular selecting the contents of SiO2, TiO2, CaO and MgO and selecting the cupola furnace, which is associated with yielding Fe(2+).

1.7 It follows from the above that C5 does not directly and unambiguously disclose a melt composition as specified in claim 1.

2. Inventive step

2.1 The aim of the patent in suit is to provide man-made vitreous fibres that are stable to high temperatures, bio-soluble, and can be produced by a spinning cup method (paragraph [0018]).

2.2 C5 relates to the similar general purpose of producing man-made vitreous fibres which are durable in use but which can be shown to be biologically advantageous (page 1, lines 3-5).

Composition "N" of C5 has most features in common with the claimed composition and constitutes the most promising springboard. This composition is consistent with the general teaching of document C5. It is irrelevant that other compositions exhibit an even higher sintering temperature, biodegradability or weathering resistance and might be even more preferred from the perspective of C5. Composition "N" of C5 constitutes the most suitable starting point for assessing inventive step.

2.3 As indicated, the patent in suit addresses the technical problem of providing high-temperature stability, bio-solubility and suitability for fiberisation using a spinning cup (paragraphs [0018] and [0056]-[0058]). Fiberisation using a spinning cup requires in particular a low liquidus temperature (paragraph [0084]).

2.4 The claimed melt composition is proposed as the solution to this technical problem, having an Fe2O3 content of 5-9 weight% and in which the proportion of Fe(2+) is greater than 80% based on total Fe.

2.5 However, the fibres of sample N in C5 are already bio-soluble (table on page 11; tested at pH 4.5, as in the impugned patent; see page 3, lines 24-26 and page 8, lines 26-30 of C5 and paragraph [0096] of the patent in suit). They are stable at high temperatures, as shown by a sintering temperature of > 900°C (C5, table on page 11). The melts in C5 are also taught as being suitable for use in a spinning cup method, with this method being specifically mentioned in C5 (page 8, lines 18-21). Even though this is not the focus of C5 and the examples are carried out using a cascade spinning process (page 10, line 8), the skilled person has no reason to disregard the explicit reference to a spinning cup method in C5.

Furthermore, the examples in the patent in suit itself provide support for the fact that the high proportion of Fe(2+) is not essential for using the spinning cup process. The comparative melt composition, having a low proportion of Fe(2+) of only 21%, could also be spun using a spinning cup.

It is merely taught that the iron oxide content, in conjunction with the MgO content of 5-7 weight%, provides good temperature stability (paragraph [0040]). The effect in comparison with higher iron oxide contents is not specifically addressed, nor is it shown to be linked to the liquidus temperature or suitability for the spinning cup process.

What the examples in the patent in suit demonstrate is that the higher proportion of Fe(2+) leads to an improvement in the high-temperature stability (paragraph [0095]).

2.6 In light of the above, the objective technical problem may be considered that of providing a melt composition having improved high-temperature stability.

2.7 It follows from the above that this technical problem is solved by the claimed melt composition, in which the proportion of Fe(2+) is greater than 80% based on total iron. There is no indication that lowering the iron oxide content would additionally contribute to this effect.

2.8 The skilled person faced with the problem of improving the high-temperature stability would be aware, for instance, of C9, which also relates to making a mineral melt which can then be fiberised to form mineral fibres (pargaraph [0001]). According to C9, cupola furnaces which have a highly reducing atmosphere have traditionally been used to make mineral melts, and almost all the iron in melts produced by cupola furnaces is in the form of Fe(2+). C9 furthermore states that Fe(2+) is good for the fire-resistance properties of the fibres (paragraph [0059]).

This teaching is not linked to any specific composition of the melt. Its general nature is also clear since similar teaching is present in C19 (page 2, lines 1-8), according to which Fe(2+) has an advantageous effect on the temperature stability of the fibres, and in C12, according to which Fe(2+) enhances the chemical and mechanical properties of the melt (page 5, lines 23-31) and is preferably produced under reducing conditions such as in a cupola furnace (page 8, lines 5-8).

2.9 The teaching of C9 would consequently prompt the skilled person to provide an Fe(2+)-containing melt when carrying out the teaching of C5, in particular as this can be achieved using a cupola furnace and C5 explicitly mentions that kind of furnace (page 8, lines 10-12). The skilled person would thus readily provide a proportion of Fe(2+) within the claimed range.

2.10 This assessment would not change even if it was additionally taken into consideration that a high proportion of Fe(2+) reduced the liquidus temperature, as stated in the patent in suit (paragraph [0044]) though not proven by experimental data. This effect would merely be a bonus effect.

2.11 As indicated, the other distinguishing feature, namely the lower iron oxide content, does not provide any technical effect. The associated partial technical problem is thus providing an alternative.

2.12 The board does not agree that C5 would teach away from lowering the iron oxide content because doing so would impair the temperature stability.

Lower iron oxide contents are consistent with the disclosure of C5 (see the table on page 11 and claim 1). C5 teaches that the amount of iron oxide (expressed as FeO) is most preferably below 8 weight%, often between 5-7 weight% (page 5, lines 32-35). According to C5, "good" fire resistance is provided by a sintering temperature of at least 800°C (page 7, lines 18-21), which is obtained in all the examples even at lower FeO contents (table on page 11). The "improved" fire resistance is even associated with an FeO content of at least 6% (page 6, lines 27-31), which is lower than the iron oxide content of sample N, although higher amounts are preferred.

2.13 The skilled person faced with the problem of providing an alternative would thus readily select a lower iron oxide content within the general disclosure of C5, for instance 8 weight% expressed as FeO (8.8 weight% Fe2O3), without the need to exercise inventive skill.

2.14 In summary, the subject-matter of claim 1 lacks an inventive step in view of C5 in combination with C9.

Auxiliary request 2

3. Claims 1 and 3 have been amended by limiting the ranges for Na2O and K2O to the respective preferred ranges, which are specified in claim 1 as granted. The basis for dependent claim 6 can be found on page 12, lines 17 to 20; the claim has been limited to ratios which are compatible with the part of the subject-matter of the independent claim to which it refers back (claim 3). There were no objections under Article 123 EPC or Article 84 EPC.

4. Sufficiency of disclosure

4.1 Appellant 3 was of the opinion that the skilled person would be unable to carry out the invention and in particular would be unable to obtain the required proportion of Fe(2+) or the Fe(0) content stipulated in dependent claims 2 and 5, referring to an Fe(0) content in the melt of less than 900 ppm, measured using a magnetic analyser.

4.2 A successful objection of lack of sufficiency of disclosure presupposes that there are serious doubts, substantiated by verifiable facts (Case Law of the Boards of Appeal of the EPO, 9th edition, 2019, II.C.9).

4.3 The patent in suit describes suitable raw materials (paragraph [0066]) and indicates melting processes (paragraph [0068] and [0076]-[0077]). It is stated that the level of Fe(2+) in the melt is increased by using a submerged arc furnace with graphite electrodes (paragraph [0068]). In an alternative embodiment, a circulating combustion chamber is used (paragraph [0076]). It is taught that when this method is used under correctly controlled conditions in terms of the redox state, the melt generally has the required proportion of Fe(2+) from the outset (paragraph [0078]). The patent in suit also gives the skilled person instructions on how to measure the amounts of Fe(2+), Fe(3+) and Fe(0) (paragraphs [0041], [0045] and [0046]), i.e. how to verify whether the desired proportion of Fe(2+) is obtained. Furthermore, it teaches that the melt and the fibres show identical ratios of Fe(2+) to total iron (paragraphs [0062], [0064] and [0065]).

4.4 The melting processes in general are known in the art. There is no proof that the skilled person would be unable to carry them out in such a way as to obtain the desired proportion of Fe(2+). In particular, appellant 3 did not attempt to rework the invention.

4.5 Appellant 3's objection concerned in particular cases in which the furnace was neither an electrical furnace nor a circulating combustion chamber. However, the patent in suit teaches that the redox state of the melt may be adjusted by, for instance, subjecting the melt to an electric potential, preferably using graphite electrodes (paragraphs [0080] and [0081]), so it is unnecessary for the melting step to directly result in the required proportion of Fe(2+). This is reflected in method claim 9.

Claims 1 and 3 relate to products (a melt composition; man-made vitreous fibres). It is irrelevant for the sufficiency of disclosure of the claimed products whether they may alternatively be produced by other methods not explicitly mentioned in the impugned patent, as long as at least one way of obtaining the claimed products is provided.

4.6 The patent also teaches that the level of Fe(0) is very low when a submerged arc furnace is used (paragraphs [0070]-[0072]). There is no evidence that this would not be the case, or that the low level of Fe(0) stipulated in claims 2 and 5 could not be obtained. A measuring method is described in paragraph [0031].

4.7 There is no reason to question the indication in the patent in suit that the redox state in the final fibres is kept as in the melt (paragraphs [0062] and [0064]), so obtaining the high proportion of Fe(2+) and the low level of Fe(0) is not linked to using - or not using - a specific spinning method.

4.8 In conclusion, the objection of insufficiency of disclosure is not convincing.

5. Novelty

5.1 Document C2 (prior art under Article 54(3) EPC)

5.1.1 The general disclosure of C2 (table on page 11) encompasses the ranges stipulated in claim 1 at issue, but the ranges known from C2 are broader. The indicated table in C2 does not disclose any specific value falling within any of the claimed ranges concerning the amounts of at least SiO2, Al2O3, TiO2, Fe2O3, MgO, Na2O and K2O. Several selections of narrower sub-ranges would thus be necessary to arrive at a composition within the scope of claim 1 at issue. C2 does not disclose limiting R2O, i.e. the total amount of Na2O and K2O, to 10 weight% or less.

5.1.2 C2 furthermore discloses three specific examples (A-C). However, in all these examples, the Al2O3 content is greater than the range stipulated in claim 1 at issue while the MgO content is lower.

5.1.3 The examples do not illustrate the most preferred range of the MgO content, which is more than 4.5 weight% to less than 6.5 weight%. An MgO content of this level (higher than 3 weight%) is linked to an Al2O3 content which is more preferably not higher than 23 weight% (page 10, line 14 - page 11, line 2).

5.1.4 Nevertheless, the teaching regarding the most preferred range of the MgO content merely identifies preferred sub-ranges in the general description of C2, i.e. in the table on page 11, namely a preferred MgO content of 4.5 weight% to less than 6.5 weight% and an Al2O3 content of 15-23 weight%. Even on this basis, it is still necessary to select sub-ranges (of at least SiO2, Al2O3, Fe2O3, TiO2, Na2O and K2O) to arrive at a composition as per the claim at issue.

5.1.5 This teaching regarding the most preferred ranges of the MgO and Al2O3 contents cannot be construed as the specific disclosure of a variant of Examples A-C in which the MgO and Al2O3 contents are modified accordingly while the remainder of the composition is maintained exactly as in these examples.

An example constitutes a specific embodiment that cannot be freely combined with other information selected within the description (T 210/05, Reasons 2.3).

5.1.6 Furthermore, even if the teaching regarding the most preferred ranges of the MgO and Al2O3 contents (page 10, line 14 - page 11, line 2) were combined with the examples in C2, this would not inevitably result in a composition within the scope of the claim because it is still necessary to also select the Al2O3 content. Moreover, the sum of the K2O and Na2O contents (R2O) is greater than required in the claim.

5.1.7 In summary, C2 does not directly and unambiguously disclose a composition within the scope of claim 1.

5.2 C1

5.2.1 The compositions disclosed in C1 overlap with those stipulated in claim 1 at issue. Considering the most relevant disclosure, namely the narrowest ranges of those disclosed in C1 (page 4), the upper limit of the SiO2 content (43 weight%), the upper limit of the Al2O3 content (23 weight%), the lower limit of the TiO2 content (1 weight%), the lower limit of the Fe2O3 content (5 weight% FeO), the lower limit of the CaO content (10 weight%), the lower limit of the MgO content (5 weight%) and the lower limits of the P2O5 and MnO contents (1 weight% each) all fall within the claimed ranges, while the other limits of all of these ranges are outside the claimed ranges. To arrive at a composition within the scope of the claim, it is thus necessary to combine upper limits (i.e. the upper parts of the range) of some components with lower limits (i.e. the lower parts of the range) of other components. These kinds of combinations of the end points of separate ranges are generally not considered disclosed (see T 1634/13, Reasons 3.2). It is furthermore necessary to select an Na2O content of 2-7 weight% and a K2O content of 3-7 weight%, but the general disclosure in C1 only concerns the sum of the two, with Na2O + K2O being at least 1 and not more than 10 weight% in C1. None of the examples provided in C1 falls within the scope of the claim at issue. Examples 5 and 6, which, it was argued, were particularly close to the claimed subject-matter, relate to MgO contents of 8.7 and 8.5 weight% and K2O contents of 0.9 and 1.6 weight%, respectively - i.e. outside the claimed ranges of 5-7 weight% MgO and 3-7 weight% K2O.

5.2.2 Appellant 1 argued that the concept developed in T 26/85 and T 666/89 had to be applied, i.e. assessing whether the skilled person would seriously contemplate working in the overlapping area. In appellant 1's opinion, this was to be distinguished from the considerations relevant to selections from lists, with a list of discrete elements in principle being different from a range of values, as is under consideration here.

5.2.3 The board does not agree that the skilled person would seriously contemplate working in the overlapping area. In this case, and as indicated above, it would be necessary to combine parts of several ranges to arrive at subject-matter within the overlapping area. This amounts to making multiple selections. The mere consideration that the ranges of a multiple selection are as such not excluded from the prior art does not establish lack of novelty (T 1095/18, Reasons 1.8). On the contrary, there is no pointer towards the overlapping area. In particular, C1 does not contain any disclosure of making the necessary adjustments to the MgO and K2O contents in Examples 5 and 6, which, it was argued, were the closest to the claimed subject-matter.

5.2.4 This case is not comparable with T 440/04, which related to the question of whether a disclaimer could be allowed. It was found that disclaiming a specific example was insufficient because the document under consideration in T 440/04 also made available fibres having compositions that were substantially equal or close to the composition of the example.

Furthermore, assessing whether an example makes the region immediately surrounding it available, as was done in T 440/04, differs from assessing whether that example makes a specific neighbouring embodiment available - or a defined part of the range surrounding it.

5.2.5 In conclusion, C1 does not directly and unambiguously disclose a composition within the scope of claim 1.

5.3 C3

5.3.1 The objection concerns the commercial mineral fibres Roxul 1000 (page 5, Figure 3). Not only are several selections necessary (the MgO content, the TiO2 content, the CaO content and the Na2O content) to arrive at a composition within the scope of claim 1, but the range of the K2O content specified in the claim at issue (3-7 weight%) is greater than what is disclosed in C3 (0-2 weight%).

5.3.2 C3 therefore does not directly and unambiguously disclose a composition within the scope of the claim at issue.

5.4 C4

5.4.1 None of the examples provided in C4 falls within the scope of claim 1. Considering the general disclosure of C4, it is necessary to select specifically the upper end value of the range of the MgO content (5 weight%) in conjunction with the upper part of the Al2O3 range (22 weight%; col. 3, lines 14-16) and the lower limit of R2O (10 weight%; col. 3, line 29). There is no pointer towards this combination.

5.4.2 C4 therefore does not directly and unambiguously disclose a composition within the scope of the claim at issue.

5.5 C5

5.5.1 C5 does not take away novelty for the reasons set out in relation to the main request (see point 1.).

6. Inventive step starting from C5

6.1 Inventive step was assessed starting from composition "N" disclosed in C5, as in the case of the main request (point 2.). The subjective technical problem is as set out in relation to the main request (point 2.3).

6.2 The claimed composition is proposed as the solution to this technical problem, and additionally differs from that known from C5 in that the Na2O content is 2-7 weight% and the K2O content is 3-7 weight%, versus an Na2O content of 0.6 weight% and a K2O content of 1.4 weight% in composition "N".

6.3 According to the teaching of the patent in suit, the alkali metal oxide helps to decrease the liquidus temperature (paragraph [0052]).

Even though there is no direct experimental evidence demonstrating this effect of alkali metal oxide, comparing the liquidus temperature generally disclosed in the patent in suit (namely less than 1 220°C or even less than 1 150°C; paragraph [0084]) with the liquidus temperature disclosed in C5 (namely 1 200°C to 1 400°C, preferably 1 240°C to 1 340°C; page 7, lines 22-24) supports the conclusion that the impugned patent relates to compositions that have a relatively low liquidus temperature. There is no reason to assume that the melt composition exemplified in the patent would not comply with the patent's general teaching and would not exhibit a liquidus temperature within the desired range. This also applies to the claimed melt compositions, which constitute a reasonable generalisation of the example. Furthermore, the appellants provided no evidence to the contrary.

In the example, good high-temperature stability was observed (paragraph [0095]).

6.4 It is therefore accepted that the technical problem of providing a balance between good high-temperature stability and low liquidus temperature, making the composition particularly suitable for the spinning cup process, is solved.

6.5 C5 does not provide any specific instructions on how to control the liquidus temperature, and in particular does not teach any benefits related to increasing the alkali metal oxide content. By contrast, low alkali metal oxide contents are preferred according to C5, the combined amount of alkali (Na2O + K2O) usually being below 5% and preferably below 3% (page 6, lines 1-3).

C5 merely teaches that increasing the amount of alkali metal oxide - among other possible measures - may increase the melt viscosity if it is too low (paragraph bridging pages 4-5). However, there is no indication that the melt viscosity of composition "N", which is higher than in the case of several other examples, would be considered "low" from the perspective of C5.

6.6 There is thus no convincing argument why the skilled person, starting out from composition "N" of C5 and faced with the technical problem of providing a balance between good high-temperature stability and low liquidus temperature, would be motivated to adjust the alkali metal oxide content to values as specified in the claim, in conjunction with the other necessary modifications discussed in relation to the main request.

6.7 For these reasons, the subject-matter of claim 1 involves an inventive step.

6.8 The same conclusion applies to claim 3, which relates to the corresponding man-made vitreous fibres having the same composition.

6.9 The same conclusion also applies to method claims 8 and 9, which relate to a method of forming fibres using the melt composition according to claim 1, and to a method of forming the melt composition, respectively.

6.10 The other claims also directly or indirectly refer back to claims 1 or 3.

7. Articles 12(2) and (4) RPBA 2007

7.1 In the statement of grounds of appeal, appellant 3 made a sweeping reference to their submissions during the opposition proceedings (filed as annexes A and B) and additionally stated that any one of C1-C4 could alternatively be regarded as the closest prior art for assessing inventive step (see page 2, first paragraph and page 11, paragraphs II.3 and II.3.1 of the statement of grounds of appeal).

7.2 In its communication pursuant to Article 15(1) RPBA 2020, the board informed the parties of its provisional opinion that this did not constitute a substantiated submission in the current appeal proceedings, and that C2 was in any case not prior art under Article 54(2) EPC.

A mere reference to a party's earlier submissions without actually dealing with, or entering into a discussion of, the reasons given in the decision under appeal by the opposition division for arriving at its decision is not enough to substantiate a ground of appeal (T 1311/13, Reasons 19).

7.3 During the oral proceedings before the board, appellant 3 stated that they maintained these objections raised in the statement of grounds of appeal with regard to auxiliary request 2. However, they did not provide any argument why the board's provisional opinion had been incorrect but explicitly acknowledged it.

7.4 The board therefore has no reason to deviate from its provisional opinion. The above-mentioned objection (see point 7.1) is therefore not taken into account (Article 12(2) and (4) RPBA 2007 in conjunction with Article 25(2) RPBA 2020).

8. Article 13(2) RPBA 2020

8.1 Appellant 2's objection of lack of inventive step in view of the combination of C5 and C7 was only raised on 17 June 2022 in reply to the board's preliminary opinion, after summons to oral proceedings had been notified.

Thus, the provisions of Article 13(2) RPBA 2020 apply. Pursuant to this article, any amendment to a party's appeal case after notification of a summons to oral proceedings shall, in principle, not be taken into account unless there are exceptional circumstances, which have been justified with cogent reasons by the party concerned.

8.2 The auxiliary requests under consideration were filed on 28 January 2019 with the patent proprietor's reply to the appeals. It would thus have been appropriate to raise any objections against them as a reaction to their being filed, without waiting for the board's preliminary opinion (see T 2843/19, Reasons 3.3).

8.3 Appellant 2 provided no reasons justifying the late filing of this objection, relying for the first time in the appeal proceedings on document C7. It is insufficient that C7 had been cited during opposition proceedings.

8.4 The objection of lack of inventive step in view of the combination of C5 with C7 is thus not taken into account.