I. The appellant (opponent) lodged an appeal against the decision of the opposition division rejecting the opposition against European patent No. 1 948 574. The patent in suit concerns crush resistant latex topcoat compositions for fibre cement substrates.

II. In the notice of opposition the grounds for opposition with respect to novelty and inventive step were raised (Article 100(a) in conjunction with Articles 52(1), 54 and 56 EPC).

III. At the oral proceedings before the opposition division, the appellant (the then opponent) stated that it did not intend to raise an objection under Article 100(b) EPC (see minutes, section 1.1).

IV. In the impugned decision, the opposition division held that the grounds for opposition mentioned in Article 100(a) in conjunction with Articles 52(1) and 54 and 56 EPC did not prejudice the maintenance of the patent as granted, having regard inter alia to the following documents:

D1: EP 0 623 659 A2

D3: WO 03/031526 A1

D4: EP 0 894 780 A1.

It rejected

D11: Celanese "Brilliant Aspects Product Range

Mowilith**(®) Emulsions for Industrial Coatings"

as inadmissible. This document was said to be late filed and prima facie not relevant because there was no direct and unambiguous disclosure of an "unattached" board.

V. With its grounds of appeal, the appellant submitted the following document:

D20: Test report ("Versuchsbericht").

VI. At the oral proceedings before the board, the appellant withdrew its request to remit the case to the department of first instance in case D20 was admitted by the board.

VII. Independent claims 1 and 17 of the main request (patent as granted) read as follows:

"1. A coated fiber cement article comprising an unattached fiber cement board substrate having a first major surface at least a portion of which is covered with a crush resistant final topcoat composition comprising a multistage latex polymer.

17. A method of making a crush resistant coated fiber cement article, which method comprises:

providing an unattached fiber cement board substrate having a first major surface;

providing a topcoat coating composition comprising a multistage latex polymer;

applying the topcoat coating composition to at least a portion of the first major surface;

drying or otherwise hardening the coating composition to form a crush resistant final topcoat; and

stacking two or more of the thus-coated boards on a pallet or other horizontal supporting surface."

VIII. The appellant's arguments, as far as relevant for the present decision, may be summarised as follows:

Admissibility of the ground for opposition pursuant to Article 100(b) EPC

This ground was not raised during the first-instance proceedings before the opposition division, a fact that was explicitly confirmed at the oral proceedings before the opposition division. Nevertheless, the ground for opposition of lack of sufficiency of disclosure should be admitted because its introduction into the appeal proceedings was justified by the submissions made by the respondent.

Admissibility of D11

Document D11 should have been admitted by the opposition division. The opposition division wrongly exercised its discretionary power because it neither took into account the breadth of claim 1 nor the problem effectively solved by its subject-matter. Moreover, it did not consider the fact that it was obvious to apply the product disclosed in D11 to unattached fibre boards. The fact that it had been difficult for the appellant to retrieve D11 was also not taken into account. As there was no improvement in crush resistance over the whole ambit of claim 1, the fact that D11 did not deal with crush resistance was immaterial for the question of prima facie relevance of

this document.

Novelty

D11 was novelty-destroying for the subject-matter of claim 1.

Admissibility of the tests submitted with the grounds of appeal (D20)

The tests submitted with the grounds of appeal were occasioned by the discussion at the oral proceedings before the opposition division and by its written decision. These tests therefore could not have been filed earlier, i.e. during the proceedings before the opposition division. There was therefore no reason to exclude these tests from the appeal proceedings.

Inventive step

D4 represented the closest prior art. More specifically, comparative examples V6 and V8 making use of a single-stage latex were to be considered the most promising starting point for assessing inventive step. The subject-matter of claim 1 differed therefrom by the presence of a multistage latex polymer in the topcoat. The problem of improving crush resistance was not solved over the whole ambit of claim 1; this was particularly evidenced by D20. Therefore, the problem to be solved was to provide an alternative coated fibre article. It was known in the art prior to the priority date of the patent in suit to use multistage latex polymers in topcoats, as evidenced by D1 and D3. It was therefore obvious to use a multistage latex polymer in the comparative examples V6 and V8 of D4, thus arriving at the subject-matter of claim 1.

IX. The respondent's arguments, as far as relevant to the present decision, may be summarised as follows:

Sufficiency of disclosure, admissibility of D11 and D20

The ground for opposition under Article 100(b) EPC was a fresh one and no consent was given to its introduction into the appeal proceedings. Nor should D11 and D20 be admitted into the proceedings. In particular, D20 was not prima facie relevant.

Novelty and inventive step

The subject-matter of claim 1 was novel. While D4 as a whole could be considered to represent the closest prior art, comparative examples V6 and V8 were not suitable starting points for assessing inventive step because they were said to perform poorly in terms of maintenance of factory appearance after application of pressure. As to the tests carried out in D20, these did not undermine the credibility of the assertion that the claimed subject-matter resulted in improved crush resistance over essentially the whole ambit of claim 1.

X. Requests

The appellant requested that the decision under appeal be set aside and that the patent be revoked.

The respondent requested that the appeal be dismissed. In the alternative, it requested that the patent be maintained in amended form on the basis of 24 auxiliary requests, submitted with the reply to the grounds of appeal.

Reasons for the Decision

1. Admissibility of the ground for opposition under Article 100(b) EPC

It is uncontested that this ground for opposition was not raised by the appellant or introduced ex officio by the opposition division in the first-instance proceedings. In particular, as can be taken from the minutes of the oral proceedings before the opposition division (section 1.1) and the impugned decision, reasons, section A, the appellant confirmed that it did not intend to raise such an objection in these proceedings. As a consequence, this ground is a fresh ground for opposition in the sense of G 10/91 (see point 18 of the reasons) whose introduction into the appeal proceedings requires the proprietor's agreement. Whether the raising of this ground of opposition was occasioned by submissions of the respondent, as contended by the appellant, is not decisive.

As the respondent (proprietor) does not agree to the introduction of this fresh ground into the appeal proceedings, the objection of insufficiency of disclosure is inadmissible.

2. D11 - admissibility

2.1 This document was uncontestedly late filed during the first-instance proceedings. Its introduction was therefore at the discretion of the opposition division (Article 114(2) EPC).

2.2 In its decision, the opposition division held that D11 was not prima facie relevant either for assessing novelty or for assessing inventive step (see section B, point 5 of the reasons). D11 did not directly and unambiguously disclose the "unattached" feature of claim 1 and was also not directed to the problem of crush resistance.

2.3 The board notes that the opposition division applied the right principle, i.e. prima facie relevance (see the decisions cited in Case Law of the Boards of Appeal (CLBA), 8th ed. 2016, IV.C.1.1.4). The opposition division also took all the relevant factors into account, i.e. any having a potential bearing on the outcome of the assessment of novelty or inventive step.

According to the appellant, the opposition division should have taken into account the alleged breadth of claim 1 and the problem effectively solved by its subject-matter. The board however considers that these arguments relate to the examination of the claimed subject-matter as to its substance and do not relate to the assessment of the prima facie relevance of a piece of evidence. Also, the difficulty the appellant reportedly had in retrieving D11 is normally not a factor to be taken into account when considering the admissibility of a late-filed document.

Thus the opposition division has not exercised its discretionary power in an unreasonable way and therefore has not exceeded the proper limits thereof (cf. the decisions cited in CLBA supra, IV.C.1.2.2a)).

2.4 The fact that the opposition division did not admit a late-filed document and did not exceed the proper limits of its discretion by not admitting it does, in principle, not prevent the board from admitting the document pursuant to Article 12(4) RPBA, in particular if it considers it to be prima facie relevant and taking into account additional facts and different circumstances (cf. T 971/11, reasons 1.2 and 1.3, and T 1811/13, reasons 7).

2.5 In the case at hand, there are no such additional facts and different circumstances such as the filing, at the beginning of the appeal proceedings, of additional evidence and of additional submissions based on the non-admitted document and the additional evidence (cf. T 971/11 supra, reasons 1.4). Moreover, D11 is considered not to be prima facie relevant. Firstly, it is not prima facie apparent that it discloses an "unattached" fibre cement board as required in granted claim 1. Secondly, it is uncontestedly unrelated to crush resistance. The question of whether D11 could be relevant because the effect allegedly did not occur over the whole scope claimed concerns an in-depth analysis of inventive step and goes beyond the assessment of prima facie relevance.

2.6 For the above reasons, D11 is not admitted into the appeal proceedings.

3. Test report D20 - admittance

D20 was filed with the grounds of appeal (cf. Article 12(1)(a) RPBA). The opposition division held that no proof was provided for the appellant's contention that for certain multistage latex polymers crush resistance was not achieved (see impugned decision, page 11, lines 4 and 5). D20 uncontestedly contains specific crush resistance tests using a number of different multistage latex polymers. Thus D20 can be considered to have been filed in reaction to the first-instance decision rejecting as inadmissible experimental data provided by the appellant during the opposition proceedings. By the same token, D20 is also considered to be prima facie relevant for the outcome of the appeal proceedings.

Therefore, there is no reason not to admit D20 into the proceedings.

4. Main request (patent as granted) - novelty

D11 is the sole document relied upon by the appellant to argue lack of novelty. Since the board did not admit this document into the proceedings, the requirement of novelty is met (Article 54(1) and (2) EPC).

5. Main request (patent as granted) - inventive step

5.1 The patent concerns prefinished fibre cement sidings (paragraph [0001]).

5.2 While the parties agree that document D4 could serve the purpose of closest prior art in principle, the respondent is of the opinion that comparative examples V6 and V8 constituted an inappropriate starting point for assessing inventive step in view of their poor performance when exposed to prolonged pressure (cf. D4, paragraph [0114] and Table 2). The question of whether examples V6 and V8 can effectively serve as the closest prior art need however not be addressed because, even starting from these examples, the subject-matter of claim 1 was not obvious in view of the prior art (see infra).

5.3 According to the patent in suit, the problem consisted of providing a prefinished fibre cement siding product that maintains its factory appearance during storage in multiple pallet stacks (paragraph [0005]). In particular, the problem solved consisted in improving crush resistance (cf. paragraphs [0022], [0087], [0089], Tables 2 and 3).

5.4 The patent proposes to solve this problem by a coated fibre cement article comprising an unattached fibre cement board substrate having a first major surface, a portion of which is covered with a final topcoat composition characterised by the final topcoat composition comprising a multistage latex polymer.

5.5 The main contentious issue between the parties throughout the written and oral proceedings was whether it was credible that the proposed solution effectively solved the problem over essentially the whole ambit of claim 1.

5.5.1 In the patent in suit four different multistage latex polymers (see examples 1 to 4 and 8 to 11; Table 2) are compared with three different single stage latex polymers (see examples 5 to 7 and 12 to 14; Table 3). It is concluded that, in terms of visual appearance (cf. Table 1), those examples using a multistage latex polymer were superior to those using a single stage latex polymer.

The comparative experiments contained in the patent in suit can be said to correspond to the closest prior art, i.e. comparative examples V6 and V8 of D4, wherein a single stage latex polymer is applied (paragraphs [0103], [0104], [0106] and [0107] of D4).

5.5.2 According to the appellant the feature distinguishing the claimed subject-matter over the closest prior art, i.e. the feature "multistage latex polymer", was broad and was only defined by a process. Also, the difference over the prior art was minimal and there was even an overlap between definition of single stage and multistage latex polymers. The examples contained in the patent were only a few specific ones in order to show that the alleged effect occurred over the whole scope claimed. It was therefore prima facie unlikely that the problem was solved over the whole ambit of claim 1.

The board observes that, according to established case law, each party bears the burden of proof for the fact it alleges (CLBA, supra, III.G.5.1.1). This means that a technical problem set out in a patent is considered to be credibly solved by a claimed invention if there are no reasons to assume the contrary. In such circumstances, the onus is normally on the opponent to prove the opposite by appropriate counter-evidence or to at least provide evidence casting doubt on the alleged solution of the problem (T 596/99, reasons 7.2.9 and T 1797/09, reasons 2.7).

The fact that the feature "multistage latex polymer" is allegedly "broad" and may be considered to be defined by the process of obtaining it (cf. paragraph [0030] of the patent in suit) is not sufficient reason to assume that the proposed solution does not solve the posed problem. Likewise, it is insufficient to argue that the patent contains few specific examples when the patent contains four examples according to the invention (paragraphs [0079] to [0082]). Also the appellant's argument that the examples do not indicate the respective amounts of polymers and therefore cannot be reworked is not sufficient to convince the board that the problem is not solved because the patent contains sufficient information concerning the amounts of the polymers to be used (see paragraph [0050]). Moreover, the evidence provided by the appellant (D20) fails to show that at specific amounts of the polymers used, no improvement over the closest prior art is obtained (infra 5.5.11).

5.5.3 As to the appellant's argument that, in view of the passage in paragraph [0030], the feature "multistage latex polymer" may also encompass single stage polymers, this argument at most concerns the clarity of the claims, which is not a ground for opposition. Moreover, it is uncontested that the closest prior art does not disclose this feature and for inventive step the question needs to be answered whether it was obvious to use multistage latex polymers in the closest prior art or not.

5.5.4 According to the appellant, none of the examples in the patent differed from the comparative examples by the use of multistage latex polymers alone, so it was not clear whether the effect was due to this feature. In particular, example 3 and comparative example 7 differed in that in example 3 butyl methacrylate was used in the monomer mixtures instead of methyl methacrylate and it could not be excluded that this had an influence on crush resistance.

It is undisputed that the single stage latex polymer in comparative example 7 was prepared using butyl acrylate, methyl methacrylate, methacrylic acid and acrylic acid whereas the first monomer mixture of the multistage latex polymer of example 3 was prepared from the same monomers except that butyl methacrylate was used instead of methyl methacrylate. The main difference between example 3 and comparative example 7 resides however in the use of a second monomer mixture (leading to the multistage latex polymer). In the absence of any evidence or compelling technical argument that would suggest the contrary, there is no reason to conclude that the improvement in crush resistance (cf. example 10 in Table 2 of the patent in suit using the multistage latex polymer of example 3 compared to example 10 in Table 3 using the single stage latex polymer of example 7) is due to the use of butyl methacrylate instead of methyl methacrylate.

5.5.5 The appellant also observed that the examples in the patent in suit differed in their Tg values from those of the comparative examples. In particular, the two Tg values for each example were considerably different. This indicated that other factors such as the use of a combination of a soft stage morphology and a hard stage morphology rather than the mere change from single to multistage latex polymers played a role in the crush resistance improvement.

The appellant, apart from pointing to the two relatively remote Tg values in the examples, has not provided any compelling technical argument why, if the Tg values were less remote or even substantially identical (cf. the patent in suit, page 4, lines 5 et seq.), the problem would not be solved. In the absence of such a compelling argument and also in the absence of any evidence in this respect, i.e. showing that a multistage latex polymer with close or even substantially identical Tg values would not result in improved crush resistance compared to single stage latex polymers, the board concludes that remoteness of the Tg values in the examples in the patent do not call into question that the improvement is due to the fact that a multistage latex polymer is used.

5.5.6 Also, the fact that claim 1 does not exclude that substantial amounts of single stage latex polymers are present, as submitted by the appellant, is no bar to recognising that the problem is effectively solved. This is because the appellant has not provided any evidence that would support the allegation that in such a case the problem would not be solved.

5.5.7 In conclusion, there is no reason based on the information in the patent alone to assume that the problem is not credibly solved by the distinguishing feature, i.e. the latex polymer being a multistage latex polymer. Thus, the onus is on the appellant to prove the contrary by providing corresponding evidence (cf. 5.5.2 supra, second paragraph).

5.5.8 Turning now to D20 (test report provided by the appellant), it is common ground that this document does not disclose any test wherein a single stage latex polymer is used. Rather all dispersions used therein are multistage latex polymers. It is therefore not possible to draw a direct conclusion from the data provided in D20 as to whether the use of a multistage latex polymer compared to the use of a single stage latex polymer does not result in improved crush resistance, as submitted by the appellant. Thus, the board needs to establish whether D20 can serve as indirect evidence for casting doubt on the success of the solution and, in particular, whether the data contained in D20 can be compared to those contained in the patent in suit.

5.5.9 Considering the substrates used in the patent and in D20, it is common ground that they are not the same. In the patent "HARDIEPLANK lap siding, SELECT CEDARMILL grade, available from James Hardie Building Products, Inc." embossed substrate with a wood grain pattern is used (see paragraph [0076]) whereas in D20 embossed substrate "Eternit Facade Panel Cedral structure uncoated CS 200 available from Eternit" (page 5, section "Untersuchte Platten") is used. No evidence is available that these substrates are substantially equivalent. In this respect, the appellant's argument that the product used in the patent was not available on the market when carrying out the tests set out in D20 does not justify the lack of similarity of both substrates.

Also, the substrate in the patent is said to be "factory primed" (see paragraph [0076]). The appellant, while admitting that the substrate in D20 was not primed, argued that the primer played a minor role because of its thinness and could be disregarded when comparing the results of the patent and those of D20.

This argument is not convincing because it is technically reasonable to assume that any additional layer on the embossed surface of the substrate contributes to crush resistance. Put differently, it is technically plausible that an alleged lack of improvement in terms of crush resistance could also be due to the absence of a primer.

5.5.10 According to the appellant, the rating system in D20 (see page 7, item 5) corresponded to the rating system used in the patent (Table 1 on page 11) and therefore these data were comparable. In any event, the rating system used in D20 was more precise and objective than the one used in the patent.

The board observes that the two systems are substantially different in that the one used in the patent is a qualitative and descriptive evaluation system (cf. "no crushed peaks", "few peaks show peak flattening to width of 2mm to 4mm" etc.) whereas the one used in D20 is of a more quantitative type (percentage of "craters > 4mm"). Thus, it is questionable whether these two systems are comparable for this reason alone. Moreover, the expression "crater" used in D20 is puzzling because no such "craters" would be expected in view of the information contained in the patent where "peaks" of the embossed patterns are "crushed" or "flattened" or the grain pattern from the opposing board is embossed into the coating that is evaluated (cf. Table 1 of the patent). The appellant contends that the expression "craters" used in D20 corresponded to "peaks showing flattening" used in the patent and "craters" were not meant to refer to circular indentations only but might also refer to elongated indentations. This argument is however not persuasive because the normal understanding of "craters" is that of circular indentations and not of elongated ones.

For the above reasons, the data contained in D20 are not comparable with those contained in the patent.

5.5.11 As can be seen from the table on page 6 of D20, crush resistance in terms of the "BF" rating ("Bruchfestigkeit", i.e. crush resistance) varies considerably depending on the weight ratio of the two latex stages contained in the multistage latex polymer. For instance dispersions 5 and 6 having weight ratios of first to second stages of 50/50 and 80/20 respectively result in a BF rating of 2 and 1 respectively (cf. also dispersions 1 and 2). According to the appellant this would show that it was not credible that the problem was solved at any weight ratio, i.e. claim 1 also encompassed multistage latex polymers in a weight ratio that led to a very poor rating and thus to no improvement vis-a-vis D4.

It should however be borne in mind that D20 does not comprise any dispersion containing a single stage latex polymer, i.e. one corresponding to those used in the closest prior art D4 (i.e. comparative examples V6 and V8 thereof). While D20 indeed suggests that crush resistance appears to vary considerably with the weight ratio of the two stages of the multistage latex polymer, D20 lacks information as to whether this variation encompasses crush resistance values that would be obtained in the closest prior art. Put differently, it is not excluded and may even be considered plausible that the lowest crush resistance obtained for the panels according to claim 1 is still higher than the one that would be obtained for the closest prior art in terms of crush resistance. D20 thus fails to show that the lowest crush resistance obtained for a panel falling within the ambit of claim 1 is only as good as or even poorer than the one obtained for the closest prior art. This reasoning applies mutatis mutandis to the appellant's argument that D20 would show that the crush resistance depended to a large extent on the Tg values of the two stages of the multistage latex polymer.

5.5.12 The fact that claim 1 also encompasses flat panels, i.e. panels which are not embossed, as contended by the appellant, is also not detrimental to recognising the success of the proposed solution, because a flat panel is suitable for being stacked on top of an embossed panel and therefore the problem of crush resistance, i.e. the problem of resistance to damage of the panel's appearance, also occurs with this type of panel. Also, as submitted by the respondent, planar panels contain surface irregularities and thus the problem of crush resistance also arises when two planar panels are stacked on top of each other. For this reason, the fact that the stacking of two planar panels according to claim 1 leads to "very good" crush resistance as evidenced by D20 (see page 8, last paragraph, first sentence) by no means proves that crush resistance is not improved with respect to the prior art. Rather it indicates that the problem of improving crush resistance is also solved for this type of panel.

5.5.13 In conclusion, D20 fails to prove that the problem is not credibly solved.

5.5.14 Thus, the problem mentioned at 5.3 supra is solved and, therefore, there is no need to reformulate the problem in a less ambitious way.

5.6 As to obviousness, it is common ground that multistage latex polymers were commonly known in the prior art. According to the appellant, the solution was obvious in view of this common general knowledge, D1 or D3.

5.6.1 The fact that multistage latex polymers were commonly known in the art is in itself insufficient to show that it was obvious to arrive at the claimed solution. In order to do so it would be necessary to show that the prior art contains a pointer to the proposed solution, i.e. teaches the use of multistage latex polymers in order to improve crush resistance.

5.6.2 D1 teaches the use of multistage latex polymers in a coating in order to improve the water whitening resistance of the coating (page 2, lines 16 to 18). D1 is however silent about the problem of crush resistance. There is no indication that it was commonly known to use multistage latex polymers in order to improve crush resistance.

5.6.3 D3 discloses multistage latex polymers which can be applied to substrates such as board and concrete (see in particular claim 1 and page 11, lines 21 to 24). This document deals with the problem of blocking resistance, i.e. the resistance of coated surfaces against sticking together when stacked or placed in contact with each other under pressure. In this document, the blocking resistance is determined by the ease of pulling apart two films made of paint composition (see "Early Blocking Test", bottom of page 12). In contrast, the problem to be solved in the present case, i.e. improvement of crush resistance, relates to surface damage caused by applying pressure. D3, while mentioning the problem of surfaces sticking together when stacked or placed in contact with each other under pressure (page 1, lines 5 et seq.), is silent about the issue of surface damage due to pressure and there appears to be nothing in D3 that would indicate that the degree of ease with which two paint films can be separated corresponds to the degree of absence of such surface damage. Therefore, D3 fails to teach a solution to the problem posed.

5.6.4 The fact that D4 does not exclude the presence of multistage latex polymers, as submitted by the appellant, is also not sufficient to render the claimed subject-matter obvious because D4 does not contain a pointer to the claimed solution.

5.6.5 The prior art therefore does not teach the improvement of crush resistance by using a multistage latex polymer. It was therefore not obvious to arrive at the subject-matter of claim 1. The same reasoning applies mutatis mutandis to independent method claim 17 and the claims depending on the latter two independent claims.

5.7 Hence, the main request complies with the requirement of inventive step set forth in Article 56 EPC.

6. In conclusion, none of the grounds of appeal set forth in Article 100(a) in conjunction with Articles 54(1), (2) and 56 EPC prejudices the maintenance of the patent as granted.

Dispositif

Order