T 2349/17 06-07-2021

Reasons for the Decision

1. Article 56 EPC

1.1 Closest prior art

1.1.1 The opposition division's conclusion according to which D1 constituted a suitable document to be taken as the closest prior art is not contested by the parties.

1.1.2 During the oral proceedings before the Board, both parties agreed that example 1 of D1 read in combination with paragraph 30 of D1, which refers to the determination of the size of the polyamide domains in a bar made of the composition prepared in example 1 thereof, constituted a suitable starting point for the assessment of inventive step. The Board has no reason to deviate from that view.

1.2 Distinguishing feature(s)

1.2.1 Paragraph 30 and example 1 of D1 read as follows:

FORMULA/TABLE/GRAPHIC

FORMULA/TABLE/GRAPHIC

1.2.2 Regarding the identification of the features distinguishing the subject-matter of operative claim 1 from the compositions prepared in example 1 of D1, the compositions defined in operative claim 1 are defined both in terms of features related to the (starting) components used to prepare the compositions (polyester; polyxylylene amide) and to the (end) composition prepared therefrom (characterised in terms of domain sizes of the polyamide dispersed in the pre melt-mixed polyester resin matrix). Those compositions are further defined as being "usable for preparing articles having high transparency" and in terms of a product-by-process feature ("obtainable by") directed to an extrusion process.

1.2.3 Regarding the reading of operative claim 1, the Board fully endorses the conclusions reached in section 2 of decision T 422/13, which may be summarised as follows:

- The compositions being claimed are directed to the products in a solid state such as those obtained at the exit of the extruder and consisting of a solid polyester matrix containing dispersed polyamide domains also in a solid form, whereby the stability feature ("to obtain a ... stable dispersion") mentioned in claim 1 is the direct and inevitable consequence of the matrix and the dispersed phases being in the solid state i.e. it is an inherent feature of the compositions as defined in claim 1;

- The melt viscosity ratio specified in claim 1 is defined on the basis of the neat copolyester (and not of its reaction product with a dianhydride of a tetracarboxylic acid).

That reading of operative claim 1, which was communicated to the parties in the Board's communication (section 6.2.2), was not disputed.

1.2.4 It was also undisputed that the subject-matter of operative claim 1 at least differs from the composition prepared in example 1/paragraph 30 of D1 in that it requires the polyamide to be dispersed in the polyester resin matrix as domains having average numerical size of 80 to 100 nm and wherein more than 80% of the domains have size of 80 to 110 nm. Indeed, these features are neither explicitly disclosed in D1, nor is any information in that respect present on file. It may only be derived from paragraph 30 of D1 that the dispersed polyamide domains in the bars prepared from the composition according to example 1 of D1 had a size of less than 500 nm, which would be understood by the skilled person to imply that these domains had an average numeral size larger than 100 nm, i.e. these domains were larger than as defined in operative claim 1. That conclusion was not disputed, in particular not by the appellant at the oral proceedings before the Board.

1.2.5 Although operative claim 1 is defined using a product-by-process formulation ("obtainable by"), its subject-matter is the composition per se. In that respect, the product-by-process feature "obtainable by pre melt-mixing ... at shear rates higher than 100 s**(-1)" was not shown to imply any additional limitation to the product itself as compared to the product defined in operative claim 1 without said feature, as put forward by the appellant (sections 56-57 of the statement of grounds of appeal). In particular, no evidence in that respect was provided by the respondent although that issue was identified in the Board's communication (section 6.2.3: first paragraph). Under these circumstances, that product-by-process feature may only be seen as corresponding to the preparation conditions of the composition being claimed, which are necessary in order to obtain the domain sizes and domain size distribution features which are also specified in operative claim 1. That reading was not contested by the parties at the oral proceedings before the Board.

1.2.6 Similarly, the melt viscosity ratio between the copolyalkylene terephthalate and the polyamide specified in operative claim 1 was not shown to additionally characterise the subject-matter being claimed as compared to the product defined in operative claim 1 without said feature (see also section 6.2.3 of the Board's communication: first paragraph). Therefore, for the same reasons as for the shear rate feature (section 1.2.5), it can only be held to correspond to a preparation condition of the composition being claimed, which is necessary in order to obtain the domain size and domain size distribution features which are also specified in operative claim 1. That reading also was not contested by the parties at the oral proceedings before the Board.

1.2.7 In view of the above, the subject-matter of operative claim 1 differs from the composition prepared in example 1 and referred to in paragraph 30 of D1 in the following features:

- the polyamide is dispersed in the polyester resin matrix as domains having average numeral size from 80 to 100 nm;

- more than 80% of the (polyamide) domains have size of 80 to 110 nm (that feature is referred to as the the domain distribution hereinafter, as was done by the parties e.g. during the oral proceedings before the Board).

1.3 Problem effectively solved over the closest prior art

1.3.1 During the oral proceedings before the Board, the respondent argued that the problem effectively solved over D1 resided in the provision of a composition having improved transparency/reduced haze. In its view, the comparison of example 1 with comparative example 1 or of example 1 with example 3 of the patent in suit showed that that problem was indeed solved.

1.3.2 However, in order to be considered in the formulation of the problem effectively solved over the closest prior art an advantageous effect should be shown to be attributable to a feature distinguishing the claimed subject-matter from the closest prior art, here either the average numeral size and/or the distribution of the polyamide domains.

a) It was undisputed that no comparison between a composition according to operative claim 1 with a composition according to example 1 of D1 is on file.

Regarding the examples of the patent in suit, it was undisputed, in particular at the oral proceedings before the Board, that according to the patent in suit the domain sizes could be either determined on a cast film (paragraphs 14-15 of the patent in suit) or on a "small bar obtained by injection molding" (paragraph 41 of the patent in suit). In that respect, it was clarified at the oral proceedings before the Board that in view of the information provided in the patent in suit, the skilled person would understand that the domain sizes were measured either on an injection moulded bar in example 1 of D1 or on a cast film in comparative example 1 and in example 3 of the patent in suit. However, no evidence was provided to show that the domain sizes determined on a cast film and on a "small bar obtained by injection molding" may be fairly compared, although that issue was identified in the Board's communication (section 6.3.1.b, last sentence of the last sub-section; see also the issue addressed in section 6.1.2 of said communication, wherein it was indicated that a specific domain size and distribution thereof appeared to be associated with the specific article on which it is determined). Under these circumstances, it cannot be concluded that the examples of the patent in suit allow a fair comparison in respect of any of the distinguishing features identified in section 1.2.7 above.

Therefore, at least for that reason, no fair comparison may be made between the examples of the patent in suit relied upon by the respondent to show that the above identified distinguishing features are related to a technical effect.

b) In addition, in the absence of any information whether or not the domain size distribution feature according to operative claim 1 is satisfied in example 1 of the patent in suit, it cannot be concluded that said example 1 illustrates the subject-matter being claimed and demonstrates the achievement of an improvement in haze of the subject-matter being claimed over the closest prior art (as put forward by the respondent in view of the comparison of example 1 with comparative example 1 of the patent in suit). Also, since example 3 of the patent in suit differs from example 1 of the patent in suit at least in that a different amount of PMDA was used, examples 1 and 3 of the patent in suit do not differ solely in (at least one of) the above identified distinguishing feature(s). Therefore, the comparison of these examples of the patent in suit also cannot support an improvement related to any of these features.

1.3.3 Nevertheless, it is noted that the appellant did not dispute in its statement of grounds of appeal that it was credible that the problem formulated by the respondent was indeed solved (section 53 of the statement of grounds of appeal).

In that respect, it is in the Board's view credible, in particular in the light of the disclosures of D27 and D36, that decreasing the size of the dispersed polyamide domains in the copolyester matrix in example 1/paragraph 30 of D1 would lead to higher transparency/reduced haze. Indeed, this is what the skilled person would expect from the indication given in D27 that, in particular in PET/MXD6 polyamide blends, haze increases with the size of the dispersed domains (abstract: "The reason for this haze is the number and size of the dispersed particles"; page 1993, left hand side column above Figure 15: "However, haze naturally becomes higher in blends with larger dispersed particles ..."; page 1994, section Conclusion: "The degree of haze ... depends on the size of the dispersed particles"). Similarly, it is derivable from pages 186-187 of D36 that it is generally accepted that the clarity of blends of incompatible polymers may be improved by reducing the domain sizes of the dispersed phase, whereby the critical size for transparency is approximately 0.1 micron (bottom of page 186). Although D36 is not specifically directed at (co)polyester/polyamide compositions according to D1, the Board considers that its teaching reflects common general knowledge regarding blends of incompatible polymers, which is the case of the polymer blend prepared in example 1 of D1. The fact that the appellant's representatives may have argued differently regarding the relationship between domain size and low haze at a different stage of the proceedings, as submitted by the respondent (rejoinder to the statement of grounds of appeal: page 4, third paragraph), is, in the Board's view, not relevant, since the teaching of both D36 and D27 is unambiguous in that respect.

1.3.4 In view of the above, it is accepted that the problem effectively solved over D1 resides in the the provision of a composition having improved transparency/reduced haze.

1.4 Obviousness

1.4.1 The question remains to be answered whether the skilled person, desiring to solve the problem identified as indicated in section 1.3.4 above, would, in view of the closest prior art, possibly in combination with other prior art or with common general knowledge, have modified the disclosure of the closest prior art in such a way as to arrive at the claimed subject matter.

1.4.2 In that respect, two main issues were in dispute between the parties, the first one being whether or not the skilled person would have any motivation to prepare compositions exhibiting the domain sizes and distribution thereof as defined in operative claim 1, the second one being if, should the skilled person be willing to prepare such compositions, he would have been able to do so on the basis of the teaching of the cited prior art documents (enablement by the prior art). These two issues are discussed separately hereinafter.

1.4.3 Motivation to prepare compositions as defined in operative claim 1

a) In that respect, D1 itself discloses that compositions are prepared in which the polyamide is dispersed in the polyester matrix with domains having size of less than 1 micron, preferably in the range from 0.2 to 0.4 micron, whereby bars obtained according to example 1 have domains of less than 500 nm (0.5 micron; see paragraph 30).

Therefore, since it is generally accepted (see section 1.3.3 above, which is equally relevant here) that for blends of incompatible polymers such as the ones used in example 1 of D1 the transparency may be improved by reducing the domain sizes of the dispersed phase, the skilled person would have a good reason for aiming at reducing and controlling the domain sizes taught in paragraph 30 for the composition prepared according to example 1/paragraph 30 of D1.

b) The respondent put forward that the skilled person would have had no motivation to improve the transparency/reduce the haze of the compositions prepared according to example 1 of D1 because D1 taught that the articles made therefrom already exhibited a high clarity (paragraph 43 of D1). The same was valid in view of Figure 15 of D27, so the respondent.

However, no evidence is on file regarding the clarity/haze level of the composition prepared according to example 1 of D1. In that respect, it is agreed with the appellant that "high clarity" as disclosed in D1 is not an absolute and unambiguous requirement. Therefore, there is no reason to expect that the skilled person would have no reason to try to further improve the clarity/reduce the haze of the composition prepared in example 1 of D1.

In Figure 15 of D27, incompatible polymer blends in which the domains of the dispersed phase in the matrix are between 100 and about 200 nm (empty dots) are indicated to exhibit a haze of 0-25 %. However, these data do not allow to conclude that said compositions could not be improved in respect of haze/transparency, as put forward by the respondent. In particular, it makes no doubt that the skilled person, considering that such compositions may exhibit a haze of up to 25 %, would not have been deterred from trying to further improve the transparency/reduce the haze of these compositions.

For these reasons, the respondent's arguments did not convince.

c) The respondent argued that the skilled person would not be motivated to decrease the domain sizes taught in D1 because it was derivable from D36 that it would lead to a decrease in terms of impact strength (reply to the statement of grounds of appeal: page 6, first four paragraphs). However, considering that the impact strength does not make part of the problem effectively solved over D1, that argument is rejected.

d) In view of the above, it is concluded that the skilled person would have had good reasons to decrease and control the size of the dispersed polyamide domains in the composition according to example 1/paragraph 30 of D1.

1.4.4 Teaching in the prior art how to prepare such dispersions (enablement by the prior art)

a) In that respect, D1 contains no specific teaching how to obtain domain sizes as defined in operative claim 1 or how to reduce and/or control the domain size distribution.

b) However, it is derivable from D27 that it was known in the art that smaller sizes of the dispersed phase domains may be obtained by increasing the shear rate during melt mixing (page 1990: right column, paragraph above Figure 8; Figure 13). In that respect, screw speeds of 200 rpm and 400 rpm are explicitly taught in D27 (page 1992: paragraph above Figure 12; Figure 13), whereby the fact that such screw speeds amounted to shear rates of more than 100 s**(-1) (it was in particular put forward at the oral proceedings before the Board that this was derivable from D27: page 1992, paragraphs above Figure 12), was not contested. It was also undisputed that shear rates of higher than 100 s**(-1) were usual in the art (statement of grounds of appeal: section 61).

In addition, D27 also teaches that smaller dispersed phase domains may be obtained by increasing the viscosity ratio polyester/polyamide up to a value of about 3-4 (Figure 14, taking into account that the viscosity ratio "p" of D27 is the inverse of the viscosity ratio defined in operative claim 1 - see D27: page 1991, right hand side column, first paragraph and page 1994: third sentence of the section "Conclusion").

Therefore, on the basis of D27, the skilled person aiming at reducing the domain sizes of the composition according to example 1/paragraph 30 of D1 would contemplate working with shear rates higher than 100 s**(-1) and with a melt viscosity ratio polyester/polyamide higher than 1.5, in particular in the higher end of the range of up to 4:1 taught in D1 (paragraph 29).

d) In doing so, it can only be concluded that, following the teaching of the patent in suit itself (which uses the same preparation process as in D1, namely a premix of polyethylene terephthalate and the anhydride, followed by a comixing with the polyamide) and in the absence of any evidence to the contrary, the skilled person would inevitably achieve the domain sizes and distribution thereof specified in operative claim 1. In that respect, in the absence of any effect related to either the specific limits of the domain sizes (nothing has been shown to be specific of the values 80 and 100 nm as opposed to values smaller than in D1, but outside that range) or the distribution thereof as indicated in operative claim 1, these features can only be held to constitute arbitrary requirements, as argued by the appellant (see paragraphs 33-35 of the statement of grounds of appeal).

e) The respondent argued that D27 neither disclosed an average domain size as small as 80 to 100 nm, nor a domain size distribution as defined in claim 1. Therefore, D27 could not lead to the subject-matter being claimed, so the respondent.

Although it is correct that D27 does not disclose the features of operative claim 1 relied upon by the respondent, in particular no domain sizes smaller than 100 nm (see e.g. Figures 11 to 15), D27 is not related to exactly the same components as the ones defined in operative claim 1 and the ones used in example 1 of D1. Indeed, the matrix is polyethylene terephthalate and not a copolymer thereof and no use is made in D27 of a dianhydride component. It remains however that the blends used in D27 and in D1 are made of very similar polymer components so that the skilled person would readily contemplate applying the teaching of D27 to the compositions of example 1 of D1, which was never contested. In addition, it is noted that D27 studied the impact of varying the shear rates and of varying the melt viscosity ratio in two separate experiments (Figures 13 and 14, respectively). Under these circumstances, it is not surprising that different results in terms of domain sizes are obtained in D27 (as compared to the patent in suit). Also, it cannot be excluded that even more noticeable effects would be obtained when combining the two measures taught in D27 simultaneously (as defined in operative claim 1). In doing so, as explained above (preceding section d), it has to be expected that the domain sizes and distribution thereof specified in operative claim 1 would inevitably be obtained.

1.4.5 For these reasons, the skilled person would know from the teaching of D27 how to decrease the size of the polyamide domains in the composition prepared in example 1/paragraph 30 of D1 and to obtain domain sizes and distribution thereof as defined in operative claim 1.

1.5 In view of the above, the subject-matter of operative claim 1 is not inventive in view of D1 as the closest prior art and the main request is not allowable.

2. Since the sole request of the respondent is not allowable, the patent is to be revoked.