T 0795/92 (Catalysts/UNILEVER) of 23.4.1996

Summary of Facts and Submissions

I. European patent No. 0 092 878 based on application No. 83 200 575. 5, filed on 21 April 1983 and claiming a priority date of 23 April 1982, was granted on 16 August 1989, on the basis of 8 claims.

II. The Respondents (Opponents 1, 2, 3 and 4) filed notices of opposition on the respective dates of 9 March, 28 April, 16 May and 15 May 1990, requesting the revocation of the patent on the grounds of lack of novelty and inventive step, insufficiency of disclosure and extension of the patent beyond the application as originally filed (Articles 100(a), 100(b) and 100(c) EPC). Of the documents cited by the parties during the opposition procedure, only D3, Journal of catalysis, volume 67, 1981, pages 186 to 206, D. G. Mustard, C. H. Bartholomew, "Determination of Metal Crystallite Size and Morphology in Supported Nickel Catalysts", and D8, GB-A-1 318 528, remain relevant for the present decision.

In the course of the opposition procedure, on 1 May 1992, the Appellants (Patentees) submitted an amended set of claims. Claim 1 thereof reads as follows:

"1. A nickel upon transition alumina catalyst, containing 5 to 40% (w/w) of nickel with an active nickel surface between 152 and 250 m2/g Ni and characterized in that the degree of reduction of nickel is between 58 and 89% and the nickel crystallites have an average diameter of 1.5 to 2.5 nanometers, the transition alumina surface area being between 90 and 115 m2/g."

III. With its decision of 2 June 1992, posted on 29 June 1992, the Opposition Division revoked the patent on the ground that the catalyst defined in the said amended Claim 1 of 1 May 1992 lacked an inventive step. The Opposition Division took the view that Claim 1 met the requirements of Article 123(2) and (3) EPC, concentrating its reasoning on the lower limit of 152 m2/g for the active nickel surface and the upper limit of 2.5 nm for the average crystallite diameter. As regards sufficiency of disclosure, the benefit of the doubt was given to the Patentees. It was accepted that the hydrogen adsorption method had been used in the patent in suit to determine the average crystallite size and that the Coenen method would have been the method chosen by the skilled man for the calculation of the degree of reduction.

IV. On 24 August 1992, the Appellants lodged an appeal against this decision. They submitted an auxiliary request (now auxiliary request 1) together with their statement of grounds of appeal, on 23 October 1992. Claim 1 of auxiliary request was essentially the same as that of 1 May 1992.

The Appellants argued in favour of an inventive step and contended that there had been a substantial procedural violation since the inventive step of the process claims was denied in the appealed decision although it had not been discussed at all during the oral proceedings.

In reply to a communication of the Board questioning the allowability of certain amendments in the claims of the main request and of the auxiliary request 1, the Appellants on 12 August 1995 submitted two further auxiliary requests (auxiliary requests 2 and 3) and on 28 March 1996 by way of evidence referred to two additional documents, namely Applied Catalysis, 75, 1991, pages 193 to 223, J. W. E. Coenen, Characterization of the standard nickel/silica catalyst EuroNi-1 (hereinafter D25) and Hydrogen Effects in Catalysis, Z. Paal, P. G. Menon, New York 1988, pages 100 to 113, (D26). The independent product claims and process claims of the auxiliary requests 1, 2 and 3 each contain the upper value 89% for the degree of nickel reduction.

Oral proceedings were held on 23 April 1996, in the absence of Respondent 1 who had informed the Board on 25 October 1995 of his intention not to attend the hearing.

V. The Appellants' arguments insofar as they concern the allowability of the 89% figure in the amended claims can be summarized as follows:

The average crystallite diameter was calculated using the Coenen formula, based on the assumption of a hemispherical model. It was clear from documents D25 and D26 that only three models existed and that Examples 1 to 4 of the original application perfectly fitted with the hemispherical model. As the data supplied in the Examples 5 and 6 did not fit with any recognized model, the skilled person would have recognized reading these examples that they contained mistakes and had to be disregarded. The fact that D25 and other documents cited in this context were concerned with nickel/silica catalysts and not with nickel/alumina catalysts was not relevant since the nature of the support had no influence on the measurement itself. D25 was a post- published document, however any review of the various methods existing up to a certain date had necessarily a later date. Other methods for determining the crystallite size, such as the magnetic measurements or the electron microscopy, were indeed known, however the former method did not constitute a standard method and the latter was not used by the Appellants as they did not have the required apparatus at that time. Furthermore, Prof. Coenen was a recognized expert in this technical field, his publications and his method were well-known and generally accepted, and it was known in the profession that he was working for Unilever at that time, so that one would have assumed that his method had been used. There was also no doubt that "Applied Catalysis" formed part of the common general knowledge. In these circumstances, the skilled person who noticed that the data in the Examples 1 to 4 of the application were in line with the Coenen method, though not those of Examples 5 and 6, would have concluded that the data in the latter examples were wrong and that the Coenen method had been used throughout the application.

VI. The Respondents contended that the auxiliary requests 1, 2 and 3 were not admissible as the independent process claims did not contain certain limitations introduced during the opposition procedure although the Appellants had previously renounced to a broader scope of protection during the said procedure.

Concerning the upper limit of 89% in Claim 1 of the main request and in the claims of the auxiliary requests, it was argued that the application did not disclose that the Coenen method was appropriate for calculating the degrees of reduction. This method did not fit with the data of at least Example 5. The skilled person reading this example would have thought that methods other than the hydrogen adsorption had been used for determining the average crystallite size, for example the magnetic measurements or the electron microscopy mentioned in D8, which was a patent specification of Unilever. The 89% value rested only on the assumption that the Coenen method was used. The fact that models led only to approximate values not fully in agreement with the reality could explain that in some examples the measured values did not fit with a model. Furthermore D25 and D26 could not be taken into consideration since, on the one hand, they were post-published and, on the other hand, they related to Ni/silica and not to Ni/alumina. The nature of the support was relevant since the model to be chosen depended on the support.

VII. The Appellants requested that the decision under appeal be set aside, that a patent be maintained on the basis of the claims submitted on 1 May 1992 (main request) or on the basis of one of the auxiliary requests 1, 2 or 3 submitted on 23 October 1992 (auxiliary request 1) and 12 August 1995 (auxiliary requests 2 and 3), and that the appeal fee be reimbursed. The Respondents requested that the appeal be dismissed.

Reasons for the Decision

1. The appeal is admissible.

2. Main request

2.1 The upper limit of 89% indicated in Claim 1 of the main request for the degree of nickel reduction is not expressly mentioned in the application as filed. The question arises whether the incorporation of this value in Claim 1 extends the subject-matter of the said claim beyond the content of the application as filed (Articles 100(c) and 123(2) EPC).

An amendment extends beyond the content of the application as filed if the amended subject-matter is not directly and unambiguously derivable from the content of the original application, even when taking into account matter which is implicit to a skilled person. This requirement clearly precludes allowing an amendment if there is any doubt as to whether or not it is derivable from the original application (cf. T 0383/88 of 1 December 1992, not published in OJ EPO).

2.2 According to the Appellants the nickel surface area SNi was determined by hydrogen adsorption, the degree of reduction DR was measured by a known method, and the nickel crystallite diameter DNi was calculated from the nickel surface area and the measured degree of reduction using the Coenen formula, based on the assumption of a hemispherical model for the shape of the nickel crystallites ( DNi = DR × 4310/100 × SNi, with DNi expressed in Å). The Appellants argued that, as both the active nickel surface area and the average Ni crystallite size were given in Table 3 of the original application, the degree of reduction could be calculated back using the Coenen formula stated above, and this calculation led to the value of 89% for the second example of Table 3.

It should be noted in this context that, on the one hand, the Coenen method was not the sole known method used before the priority date to determine the average nickel crystallite size in nickel supported catalysts and, on the other hand, the various known methods do not lead to identical results (see for example D8 and D3). It follows therefrom that the allowability of the value 89% depends on whether or not it is directly and unambiguously derivable from the application as filed that the average crystallite diameter was calculated from the nickel surface area and the measured degree of reduction, using the Coenen method.

2.3 The methods used for determining the nickel surface area, the average crystallite diameter and the degree of reduction are not mentioned in the original application, let alone the formula by which one of these parameters might have been calculated. Nor does the application contain any reference to the Coenen method or to the kind of assumption made for the shape of the nickel crystallites (spherical, hemispherical or cubic model). Therefore, the average crystallite diameter might have been determined by any one of the methods known before the priority date to be appropriate for measuring the average nickel crystallite size in supported catalysts, for example (i) by magnetic measurements, (ii) by electron microscopy, or (iii) by calculation from the nickel surface area obtained by hydrogen adsorption and from the measured degree of reduction, using either the Coenen model (i.e. the hemispherical model) or another model, in particular the spherical model (cf. D3). It was not contested by the parties that these methods formed part of the common general knowledge in this technical field before the priority date. However, as neither the spherical model nor the cubic model fits with any of the data reported in the Examples 1 to 6 of the original application, the skilled person would conclude that in these examples the average crystallite size was not calculated from the hydrogen adsorption data assuming a spherical or a cubic model. Although it is disclosed in D3 that, in the case of those particular Ni/alumina catalysts, owing to the lack of contrast resulting of the fine pore structure of the support (cf. page 191, left-hand column), the measurement of the average crystallite size by transmission electron microscopy is limited to a certain range of nickel loadings, this does not mean that this method is not suitable for the catalysts of the patent in suit. Furthermore, it was not contested by the Appellants that both the magnetic measurements and the transmission electron microscopy are indeed suitable for determining the average nickel crystallite size in nickel/alumina catalysts having an average crystallite size up to 5 nm. The Appellants' arguments that the magnetic measurements did not represent a standard method do not preclude that this method might have been used in the patent in suit. It should be noted in this respect that in D8, a patent specification of Unilever, the average crystallite size of nickel supported catalysts was determined by either magnetic measurements or hydrogen adsorption or electron microscopy.

The Appellants' arguments that the Coenen model was generally accepted and that it was known in the profession that Prof. Coenen was working at Unilever at that time (cf. point V above) are not sufficient to convince the Board that it is directly and unambiguously derivable from the original application that the Coenen method was used throughout the application, for the following reasons: The values of nickel surface area, average crystallite diameter and degree of reduction reported in examples 1 to 4 of the original application do indeed fit with the Coenen formula, but this does not apply to the values given in examples 5 and 6, where the application of the Coenen formula would lead to average crystallite diameter of 3.4 nm and 3.6 nm, respectively, as opposed to the values of 3.0 and 3.8 nm actually shown. Taking into account that in two out of six examples the data do not fit with the Coenen formula, the skilled person would not necessarily infer from the original application that the Coenen method had been used throughout the application and that these two examples contain mistakes. Another plausible explanation for this discrepancy would be that the average crystallite size had not been calculated from the hydrogen adsorption data but measured by one of the other known methods mentioned above, namely the magnetic measurements or the electron microscopy, even if these methods were not as commonly used as the hydrogen adsorption method.

D25 and D26, which were published well after the priority date and upon which the Appellants relied to support their arguments, do not contain any information showing that electron microscopy and magnetisation measurements are not suitable for the determination of the average crystallite size of nickel supported catalysts.

It follows from the above that, in spite of a certain probability in favour of the Appellant's position, it is not directly and unambiguously derivable from the application as filed that the average diameter of the nickel crystallites was calculated from the hydrogen adsorption data and the measured degree of reduction, using the Coenen model. Therefore, the value 89% calculated back on the basis of the Coenen formula is itself not directly and unambiguously derivable from the original application. For this reason Claim 1 of the main request does not meet the requirements of Article 123(2) EPC, and this request must be refused.

3. Auxiliary requests 1 to 3

Each of the auxiliary requests contains an independent claim including the upper limit 89% for the degree of nickel reduction. Hence, the reasons given above apply likewise to the auxiliary requests, which must therefore be refused on the same ground as the main request. In these circumstances, it was not necessary to examine the question of admissibility of these auxiliary requests.

4. Reimbursement of the appeal fee

According to Rule 67 EPC, reimbursement of the appeal fee shall be ordered where the Board of Appeal deems an appeal to be allowable, if such reimbursement is equitable by reason of a substantial procedural violation. As in the present case the appeal is not allowable, the Board would not be in a position to order reimbursement of the appeal fee even if a substantial procedural violation had indeed occurred (which the Board cannot see to be the case).

ORDER

For these reasons it is decided that:

1. The appeal is dismissed.

2. The request for the reimbursement of the Appeal fee is refused.