T 1853/21 (Pelletted animal feed/ANITOX) of 12.1.2023

Summary of Facts and Submissions

I. The appeal was filed by the patent proprietor against the opposition division's decision to revoke the European patent.

II. With their notices of opposition, the three opponents had requested revocation of the patent in its entirety on the grounds under Article 100(a) (lack of novelty and lack of inventive step), 100(b) and 100(c) EPC.

III. The opposition division found, inter alia, that claims 1 and 4 of the main request did not contain added subject-matter and did not extend the scope of protection. However, it found that the expression "The ethoxylated castor oil surfactant has 1 to 200 ethylene oxide molecules distributed normally around a mean of 40 to 80" in claim 1 of the main and auxiliary requests 1 to 5 lacked clarity.

IV. Claim 1 of the main request reads:

"1. A method for making pelleted animal feed, comprising:

(i) preparing a stock composition containing,

a) 10-90 wt %. of a buffered or an unbuffered organic acid selected from the group consisting of acetic, propionic, butyric acid and mixtures thereof,

b) 1-90 wt.% of ethoxylated castor oil surfactant having an HLB from 4 to 18 and a molar ratio of 1 molecule of castor oil to 1 - 200 molecules of ethylene oxide;

c) 0 - 20 wt.% of antimicrobial terpenes, or essential oils;

d) 0 - 40 wt.% of water;

wherein the ethoxylated castor oil surfactant has 1 to 200 ethylene oxide molecules distributed normally around a mean of 40 to 80,

(ii) adding water to prepare a heat-treating composition and applying an effective amount of said heat-treating composition to an animal feed, with sufficient heating to pelletize or extrude the feed."

V. The documents submitted during the opposition proceedings included:

D34:|"Polyoxyl Castor Oil, Polyoxyl Hydrogenated Castor Oil, Summary Report", The European Agency for the Evaluation of Medicinal Products, June 1999|

D52:|Extract from Wikipedia: "Normal distribution" |

D54:|Ullmann's Encyclopedia of industrial Chemistry, Chapter "Surfactants", 1994, 747-817 |

D55:|Fruijtier-Pölloth C., Toxicology, Vol. 214, 2005, 1-38 |

| |

VI. With its statement setting out the grounds of appeal, the proprietor (appellant) filed a main request, corresponding to the main request of the decision under appeal, and auxiliary requests 1 to 23. Furthermore, the appellant filed amended pages 3, 7 and 8 of the description and the following documents:

D56:|M. Balcan et al., Colloid Polym Sci, Vol. 281, 2003, 143-5 |

D57:|E. H. Crook et al., The Journal of Physical Chemistry, Vol. 67(10), 1963, 1987 |

D58:|Handbook of Radioactivity Analysis, 2nd Edition, Academic Press, 2003, Chapter 8, "Sample Preparation Techniques for Liquid Scintillation Analysis" by James Thomson, 655-6 and 660|

D59:|"Ethoxylation Reactor Modelling and Design", thesis by Yen-ni Chiu submitted toSwinburne University of Technology, Melbourne, Australia, 2005, 17 |

D60:|Polymer Exemption Guidance Manual, US-EPA, 1997, 1-4, 7 and 13-14 |

VII. The arguments presented by the appellant can be summarised as follows.

- D56 to D60 were to be admitted in appeal. They were filed in reaction to the opposition division's unexpected finding of lack of clarity and to the opponent's filing of D52 to D54 during the oral proceedings held before the opposition division.

- The claims did not contain added subject-matter. The feature "0-40% of water" was based on page 9, lines 25 to 26 as filed; the feature "the ethoxylated castor oil surfactant has 1 to 200 ethylene oxide molecules distributed normally around a mean of 40 to 80" was based on page 9, lines 20 to 21 as filed.

- The feature "the ethoxylated castor oil surfactant has 1 to 200 ethylene oxide molecules distributed normally around a mean of 40 to 80" was clear because ethoxylated castor oil surfactants were well known, commercially available and defined by standard nomenclature. The "mean" in claim 1 concerned the number of ethylene oxide moieties. The person skilled in the polymer field understood that claim 1 encompassed a distribution of ethylene oxide moieties approximating a normal distribution, that not all numbers from 1 to 200 had to be represented and that in the relevant literature the terms "mean" and "average" were used interchangeably. The absence of a reference to a standard deviation did not render claim 1 unclear either.

- Should any request fulfil the requirements of Articles 84 and 123(2) EPC, the case had to be remitted to the opposition division for further prosecution.

VIII. The arguments presented by opponent 2 (respondent 2) may be summarised as follows.

- D56 to D60 were filed late and not relevant. The issues relating to the term "normal distribution" were raised during the oral proceedings held before the opposition division. Thus, D56 to D60 should have been filed during the opposition proceedings and should not be admitted into the appeal. Amended pages 7 and 8 of the description should not be admitted either.

- The two following features in claim 1 added originally undisclosed subject-matter: "0 to 40% of water" and "wherein the ethoxylated castor oil surfactant has 1 to 200 ethylene oxide molecules distributed normally around a mean of 40 to 80%".

- The expression "wherein the ethoxylated castor oil surfactant has 1 to 200 ethylene oxide molecules distributed normally around a mean of 40 to 80%" was unclear. This expression was mathematically incorrect because it required incompatible constructions of the concepts of "mean", "normal distribution" and "spread". The given values of mean and spread required the curve to be so skewed that it could not be considered "normal". The standard deviation and the method for determining the mean number of ethylene oxide moieties were also not specified.

The requests

IX. The appellant (patent proprietor) requested that the decision under appeal be set aside and that the patent be maintained on the basis of:

- the claim set according to the main request or, alternatively

- one of the claim sets according to auxiliary requests 1 to 23 (all requests filed with the statement of grounds of appeal)

The appellant further requested that the aforementioned claim sets be combined with amended pages 7 and 8 of the description (as filed with the patent proprietor's grounds of appeal), which are labelled as "Auxiliary request-Description" 1 and 2, respectively (the remaining parts of the patent specification, including page 3, being left unamended).

X. The respondent 2 (opponent 2) requested that the appeal be dismissed.

XI. Opponents 1 and 3 (respondents 1 and 3) have not taken any procedural steps in the appeal proceedings. Opponent 3 withdrew its opposition on 3 November 2022.

Reasons for the Decision

1. Amendments to claim 1

1.1 Respondent 2 considered that the incorporation into claim 1 of the two following features added originally undisclosed subject-matter:

- "0 to 40 wt.% of water"

- "wherein the ethoxylated castor oil surfactant has 1 to 200 ethylene oxide molecules distributed normally around a mean of 40 to 80"

1.2 The first feature, requiring the stock solution of claim 1 to contain 0 to 40% of water is disclosed verbatim on page 9, lines 26 to 27 of the application as filed. Respondent 2 submitted that this feature was not disclosed in combination with the other features characterising claim 1 as filed. Thus, its addition into claim 1 created new subject-matter.

1.3 This argument is not convincing. The application as filed, as a whole, teaches that the invention relates to a method for making a pelleted animal feed (see claim 1 and pages 7 and 9 of the description). The method involves the preparation of a "stock composition", which is then diluted to prepare a heat-treating composition. The heat-treating composition is used for pelletising. The stock formulation of claim 1 is identical to that shown on page 7, lines 5 to 24 and on page 9, lines 4 to 15 as filed. Pages 7 and 9 simply refer to a "composition", without mentioning the word "stock". However, it is clear that the composition described in these pages, which is then diluted to form a heat-treating composition, qualifies as a stock composition and corresponds to that of claim 1.

1.4 It is thus also clear that the feature on page 9 lines 26 to 27, "The stock composition may comprise 0 to 40% water", relates to the same stock composition described in lines 4 to 9 of page 9 and in claim 1 of the application as filed. It is also clear that this feature is not inextricably linked to the other features mentioned on this page, lines 15 to 29.

1.5 Accordingly, the addition of the wording "0 to 40% water" to claim 1 does not create new subject-matter.

1.6 The feature requiring that the "ethoxylated castor oil surfactant has 1 to 200 ethylene oxide molecules distributed normally around a mean of 40 to 80" derives from page 9, lines 20 to 21 of the application as filed.

1.7 Respondent 2 noted that the passage on page 9 refers to 1 to 200 "ethylene" and not "ethylene oxide" molecules. In its opinion, this change in wording results in a change of technical teaching and in the creation of originally undisclosed subject-matter.

1.8 This argument is not persuasive. The teaching of the application as filed must be assessed as a person skilled in the art would. Such a person knows that ethoxylated castor oil is an ethylene oxide condensate of castor oil. Its molecules comprise a fatty acid, typically ricinoleic acid, functionalised using ethylene oxide, to obtain the following structure:

FORMULA/TABLE/GRAPHIC

1.9 The length of the fatty acid is typically defined referring to the number of carbon atoms and not the number of ethylene units. Furthermore, fatty acids typically contain from 6 to 21 carbons and comprise neither 1 nor 200 ethyl residues, as specified by the aforementioned wording of claim 1.

1.10 Conversely, the length of the ethoxylated residue can vary considerably. It can contain one and typically more than one ethoxyl residue which forms a polyethylene glycol chain. This chain may contain 200 ethoxyl residues.

1.11 Therefore, the skilled person would promptly understand that the wording "1 to 200 ethylene molecules" used on line 20 of page 9 of the application refers to the ethoxylated residue attached to the castor oil molecule and not to the fatty acid moiety. Moreover, claim 1 and page 7, lines 9 to 11 and page 9, lines 7 to 8 already refer to 1 to 200 molecules of ethylene oxide.

1.12 The skilled person would therefore understand that the wording on page 9 contains an error and that the wording "ethylene molecules" should be "ethylene oxide molecules". Thus, as decided by the opposition division, the amendment in claim 1 includes the correction of an obvious error, which does not create new subject-matter.

1.13 Respondent 2 submitted that the combination of the two aforementioned features adds new subject-matter. This argument is not persuasive either.

1.14 Page 9, lines 20 to 21 of the application as filed teaches that castor oil has "1 to 200 ethylene molecules distributed normally around the mean, preferably a mean of 40 to 80". This is the only further characterisation of the castor oil in the patent application. Accordingly, the skilled person would understand that the castor oil of claim 1 has this structure or, at least, that this structure is preferred.

1.15 As mentioned above, the feature on page 9, lines 26 to 27 of the application, requiring the stock solution to comprise 0 to 40% water, relates to the stock solution of claim 1. Water is the only additional ingredient mentioned on page 9.

1.16 For these reasons, the combination of the two aforementioned features does not create added subject-matter extending beyond the content of the application as filed (Article 123(2) EPC).

2. Amendments in pages 7 and 8 of the description

2.1 Pages 7 and 8 of the description as granted contained two amendments specifying that Formulae 2a and 4b and Treatments 2, 4 and 6, shown in tables 3 and 4, were representative of the invention. During the opposition proceedings, respondent 2 argued that these amendments added originally undisclosed subject-matter. However, the opposition division did not find this argument persuasive.

2.2 With its statement of grounds of appeal, the appellant filed, as an auxiliary request, amended pages 7 and 8 in which the aforementioned amendments were deleted. In reaction to a preliminary opinion from the board that pages 7 and 8 as granted contained added subject-matter, the appellant requested, as its main request, maintenance on the basis of amended pages 7 and 8. This renders the earlier added-matter objection moot. There is no reason not to admit amended pages 7 and 8 into the appeal proceedings: these pages were filed at the very first stage of the appeal proceedings as a pre-emptive measure should the board have overruled the opposition division's finding. Furthermore, the amendments overcome the objection without introducing any new issue or negatively affecting the economy of the procedure (Article 12(4) RPBA 2020).

3. Admissibility of D56 to D60

3.1 D56 to D60 were filed by the appellant with its statement setting out the grounds of appeal to address the objection of lack of clarity which led to the opposed patent being revoked.

3.2 Respondent 2 requested that D56 to D60 not be admitted into the appeal proceedings. In its opinion, the appellant should have filed these documents during the opposition proceedings because the clarity objection had already been raised during those proceedings. Respondent 2 conceded that, after receiving the opposition division's preliminary opinion that the claimed matter was clear, it had further developed its attack and filed new documents in evidence. However, before the date scheduled for the oral proceedings, the appellant was able to react to the new submissions in writing, providing arguments and a new document (D55). Any additional document should have been filed at that stage. There was no justification to present a new case on appeal. Furthermore, D56 to D60 did not represent common general knowledge and were only excerpts of the original documents, rendering an evaluation of the overall teaching of their disclosure impossible.

3.3 These arguments are not convincing. It is not disputed that the objection of lack of clarity was raised early in the written proceedings before the opposition division. In its second communication issued in preparation for the oral proceedings, the opposition division took into account the arguments that had been provided by the parties and expressed the preliminary opinion that the claimed subject-matter was clear.

3.4 After receiving that preliminary opinion, on the very last day of the period set by the opposition division for providing further submissions, the opponent presented further arguments and submitted new documents (D52 to D54) further developing its clarity attack. The new arguments related to the mathematical interpretation of the wording of claim 1.

3.5 The proprietor tried to address the new submissions in writing and filed a new document, D55, before the date scheduled for the oral proceedings. However, during the oral proceedings, taking into account the new arguments and documents filed by respondent 2, the opposition division revoked the patent (see the decision under appeal, point 4.2 and in particular the reference to D52 and D54).

3.6 The filing of D56 to D60 is therefore a reply to the new submissions presented by respondent 2 at a late stage of the opposition proceedings, in reaction to the opposition division's positive preliminary opinion. The fact that the appellant addressed the new objections before the oral proceedings attests that it promptly attempted to rebut the new attack. However, this is not a reason for precluding the appellant from providing further arguments and documents in appeal to address late developments which resulted in the reversal of the opposition division's preliminary opinion and in the revocation of the patent.

3.7 Furthermore, D56 to D60 are highly relevant for the discussion on whether the claimed subject-matter is clear. In fact, they provide relevant information on the properties of ethoxylated polymers. Moreover, contrary to the submissions of respondent 2, they do not result in the presentation of a "new case". There is also no reason to consider D56 to D60 unsuitable for conveying the relevant information on the open issues. The submitted parts of these documents are clear and self-contained. Thus, complete versions of these documents are not necessary to confirm the relevant teaching. Finally, the parts of these documents mentioned by the appellant concern general statements considered to represent the skilled person's common general knowledge at the relevant date.

3.8 For these reasons, and considering that they were filed at the very first stage of the appeal proceedings, D56 to D60 are not to be disregarded in these proceedings (Article 12(4) RPBA 2020).

4. Clarity

4.1 The opposition division decided that the wording in claim 1 "the ethoxylated castor oil surfactant has 1 to 200 ethylene oxide molecules distributed normally around a mean of 40 to 80" was unclear.

4.2 Respondent 2 endorsed the opposition division's finding. It submitted that the lack of clarity arose because it was mathematically impossible to simultaneously have a symmetric normal distribution of the number of ethylene oxide moieties around a value in the range of 40 to 80 and a spread having values between 1 to 200.

4.3 In other words, in its opinion, it was impossible to reconcile the three requirements that:

- the mean value for the number of ethylene oxide moieties was 40 to 80

- the spread of values for the number of ethylene oxide moieties of the ethoxylated castor oil surfactant was 1 to 200

- the number of ethylene oxide molecules was distributed normally and, therefore, also symmetrically around the mean

4.4 Since the mean was not at the centre of the spread, the distribution ought to have a skewed shape, which could not be considered normal according to the mathematical meaning of this term. The degree of skewness was such that the distribution could not even be considered an approximation of a normal symmetric distribution. In addition, claim 1 did not define the standard deviation of the distribution.

4.5 Respondent 2 noted that there were cases where the different features specified in a claim were separately clear but inconsistent with each other. In this case, their combination and the claimed subject-matter were unclear. The case law did not allow illogical and contradicting features to be disregarded. If a claim included contradictory features, the contradiction could not be resolved by merely disregarding the technically inaccurate feature and considering only the convenient technically meaningful feature. In the current case, the appellant had decided to define the claimed subject-matter in mathematical terms. Therefore, it had to accept that the terms in claim 1 be accorded their standard mathematical meaning. Claim 1 was thus unclear.

4.6 These arguments are not persuasive.

4.7 As set out by the appellant, the skilled person would read claim 1 on a technical and reasonable basis, using common general knowledge and avoiding artificial and semantic interpretations. The skilled person would determine whether there is a practical meaning of this claim.

4.8 Respondent 2 gives a purely mathematical interpretation of claim 1, interpreting the range of 1 to 200 as a mathematical "spread" and assuming that all distributions having a mean of from 40 to 80 ethylene oxide moieties include values across the entire range of 1 to 200. The term "normal" in claim 1 is also assumed to define a perfectly symmetric distribution around the mean.

4.9 This interpretation is not based on a technically reasonable reading of the claim. In the current case, the skilled person is not a mathematician but rather a person working in the field of polymer chemistry.

4.10 Such a person would understand that a pure mathematical approach cannot be applied to polymer chemistry due to the highly empirical nature of this field of technology. They would also understand that it would be impossible to obtain a perfectly symmetric distribution of ethylene oxide moieties around the mean value in practical real world scenarios. This understanding is corroborated by the fact that ethoxylated polymers are known to have a skewed Poisson distribution of chain lengths, which deviates from the symmetrical bell shape of a mathematically calculated Gaussian distribution (see D54, page 786, left-hand column, referring to the Poisson distribution of the species obtained in ethoxylation reactions; D56, page 143, right-hand column, referring to the Poisson distribution of the chain lengths of polyethylene oxide surfactants; D57, page 1987, note 14 and D59, point 2.4.2.1, confirming that for polyethylene oxide derivatives, a normal distribution can be approximated by a Poisson distribution; deviations from normal distributions are also discussed in D52, a Wikipedia extract relating to "Normal distribution" see the "Approximate normality" section).

4.11 Respondent 2 submitted that these documents did not mention ethoxylated castor oil. Nevertheless, all of them, with the exclusion of D52, which only concerns mathematical issues, relate to non-ionic surfactants comprising polyethylene oxide chains. Thus, their teaching is relevant for the current case. Respondent 2 observed that the term "average" rather than "mean" is used in some of these documents. However, in the relevant field, the terms "average" and "mean" are used synonymously when referring to molecular weight (see e.g. D60, page 4).

4.12 The absence of an indication of a standard deviation in claim 1 does not result in a lack of clarity either. The standard deviation provides a measure of the variation in a set of values, i.e. whether the values are narrowly or broadly distributed (see D52). If the standard deviation is not indicated, this means that the claim encompasses all possible Gaussian curves having a mean falling within the given range. Remarkably, D56 to D58 teach that polyethylene oxide compounds, including castor oil, have polymer chain lengths in a normal distribution, but none of them mentions a standard deviation.

4.13 While respondent 2 attaches great importance to the term "spread" and its mathematical meaning, claim 1 does not even mention this term. For this reason, and considering the type of ethoxylated surfactant claimed, respondent 2's interpretation of claim 1 is not convincing. This means that claim 1 does not require all distributions of values having a mean of from 40 to 80 ethylene oxide moieties to contain values spread across the entire range of 1 to 200, in particular where the mean is close to the lower end of the range of 40 to 80.

4.14 In the written proceedings, the appellant also argued that claim 1 did not indicate how to determine the mean number of ethylene oxide moieties in the surfactant molecule. The method for measuring the average molecular weight of the surfactant, which allegedly is to be used for determining the mean number of ethylene oxide moieties, was not indicated. In its opinion, since different methods would have led to different mean values, the expression "mean of 40 to 80" was, as such, unclear.

4.15 This argument is not convincing either. In the first place, claim 1 does not mention any molecular weight but rather the number of ethylene oxide molecules in the castor oil surfactant. Furthermore, the definition of the castor oil surfactant given in claim 1 is that typically used for defining the structure of ethoxylated castor oil and other non-ionic surfactants in the relevant art, see:

- paragraph 1 of D34, a report from the EMEA, which refers to "polyoxyl n castor oil (n=30 to 40)" and explains that the number n defines the average number of ethylene oxide units in the surfactant compound

- the definitions of the "PEG castor oils" given in D55, in particular the last row of table 1 on page 6 and the last full paragraph of page 8, left-hand column

- the definitions used for defining the length of the ethylene oxide chain of non-ionic surfactants disclosed in the introduction of D57

4.16 Also, it was not disputed that the measured average molecular weight of certain polymeric compounds may be subject to differences, depending on the method used for the determination. However, all aforementioned documents - including D34, an official report issued by the European authority that evaluates medicinal and veterinary products relating to, like the opposed patent, the use of ethoxylated castor oil in veterinary applications - refer to that average number of ethylene oxide molecules without any indication of the method used for its determination. This shows that in the relevant field, the skilled person would consider the indication of this number to be sufficiently clear to define the structure of the relevant castor oil surfactant.

4.17 Furthermore, there is no evidence that uncertainty would arise for the molecules claimed. Respondent 2 illustrated a theoretical example to show that the average number of ethylene oxide moieties in an ethoxylated chain could vary depending on whether it is calculated using the number- or weight-average molecular weight of the molecule. However, this example is theoretical and concerns a molecule which differs considerably from the claimed one in terms of structure and number of ethylene residues.

4.18 Accordingly, the definition of the ethoxylated surfactant given in claim 1 is clear (see also decision T 1115/10, point 9.3 of the Reasons).

4.19 For these reasons, claim 1 does not infringe the requirements of Article 84 EPC.

5. Remittal of the case

5.1 As submitted by the appellant, there are significant, undecided issues relating to Articles 83, 54(2) and 56 EPC. These involve essential questions on the patentability of the claimed subject-matter and evidence filed by the parties which is in dispute and has not been evaluated. An objection of lack of novelty based on public prior use was also raised, relying on numerous documents, and the relevance of post-filed experimental data for inventive step is also disputed. These issues were not discussed during the oral proceedings before the opposition division, and the division refrained from discussing them in its decision.

5.2 Thus, there are special reasons for remitting this case to the opposition division (Article 11 RPBA 2020).

Order

For these reasons it is decided that:

1. The decision under appeal is set aside.

2. The case is remitted to the opposition division for further prosecution.