T 0861/98 22-11-2001

Reasons for the Decision

1. The appeals are admissible.

2. Public availability and state of the art

2.1. As the public availability of certain relevant pieces of information is in dispute in the present case, it is appropriate to start from the principles pointed out e.g. in decision G 1/92 of the Enlarged Board of Appeal, wherein it is stated (Headnote):

"The chemical composition of a product is state of the art when the product as such

- is available to the public and

- can be analysed and reproduced by a skilled person, irrespective of whether or not particular reasons can be identified for analysing the composition."

2.2. In the present case, the product in question is a transparent free-standing polycrystalline diamond film as defined in claim 1 and 4 (see point VI, above) and the point in time is the priority date of the patent in suit (14 September 1989).

2.3. In accordance with established jurisprudence, information becomes "available" to the public within the meaning of Article 54(2) EPC and thus state of the art, when only a single member of the public, i.e. a person which is under no (express or implied) obligation to secrecy regarding that information, is in a position to gain access to it in any way - e.g. by analysing a product made available to such a person.

2.4. It is evident from the correspondence and not in dispute between the parties that before the priority date of the patent in suit Mr Anthony of GE sent several samples of free-standing diamond films to Prof. Koidl of Fraunhofer for scientific analysis. A first sample was received by Prof. Koidl in 1988 (see D14a, first page, last paragraph; letter of Mr Anthony to Prof. Koidl, dated December 16, 1988 = enclosure 2 to D14a, second paragraph, last line); from its examination at Fraunhofer the basic physical data were obtained that then were disclosed orally in April 1989 in a lecture at a public conference (see D14, D14a). This first sample was characterized in Mr Anthony's aforementioned letter of 16 December 1988 as being an "unpolished opaque piece" (see D14a enclosure 2; D31c). Enclosed with this letter, Mr Antony forwarded several slices of a second "unpolished semi-transparent" sample to Fraunhofer for investigation and asked to compare these data with those obtained from the previously sent "opaque film". A further diamond strip which, when holding it against the light, exhibited a semi-transparent center between opaque opposing ends was sent to Fraunhofer enclosed with Mr Anthony's letter dated 1 March 1989 (D31d). In response to this letter, Prof. Koidl - after Raman spectroscopy, X-ray diffraction and grain size measurements - held that "transparency (of that last sample) was enhanced by the waveguide-effect of the 110-fibres" (D26 = letter dated 2. April 1989, last two paragraphs).

2.5. There is no submission nor any objective indication of an explicit agreement on confidentiality regarding the properties of or any other information related to the diamond film samples made available by GE/Mr Anthony to Fraunhofer/Prof. Koidl. Equally, construing an implicit obligation to secrecy on the part of Fraunhofer/Prof. Koidl would be in evident conflict with the prevailing circumstances and the behaviour of the persons involved, in particular the content of the correspondence exchanged between Prof. Koidl and Mr Anthony and the fact, that Mr Anthony never objected to the presentation of the results of the scientific investigation, which he had asked Prof. Koidl to carry out on the samples sent to him, at the Electrochemical Society conference to be held in May 1989. Even if Mr Anthony had not the same intention to publish from the beginning, he would have become aware thereof at the latest from Prof. Koidl's letter (fax) to him dated 24. November 1988 (D31b), to which a draft extended abstract of the planned contribution for the ECS conference was attached (D14a, enclosure 1). Mr Anthony stated in response to this letter that he "could not have improved" that abstract (D31c). Moreover he was also - certainly not without his knowledge and consent - indicated as co-author on the Extended Abstract of the lecture delivered at the ESC conference 1989, and on its complete version (D1b) "Optical and Structural Characterization of CVD Diamond". According to document D30 the abstract and the extended version were published on 29 September 1989. It fits into the picture that (before the priority date of the patent) at least two other persons who qualify as members of the public (D17, D23: Prof. Kassing; D22, D22a, D22b: Mr Matsumoto) obtained preprints of D1b from Prof. Koidl. The above circumstances including the comprehensive oral disclosure (see D1a and D1b) which had taken place in the meanwhile at the ECS conference, demonstrate that Professor Koidl obviously did not consider himself or his colleagues from Fraunhofer to be constrained by an obligation to confidentiality in any matter concerning the GE diamond film samples.

2.6. In view of the strength of the evidence submitted and considering all the relevant circumstances, the following has to be concluded: In the absence of any obligation to secrecy vis-à-vis GE/Anthony, Professor Koidl qualified as a(n expert) member of the public. This means that the GE diamond films, their physical properties including those actually established by the analysis at Fraunhofer and those which could have been extracted by means of known analytical techniques which were available to a skilled person like Prof. Koidl before the relevant filing date (decision T 952/92), as well as all the technical information Mr Anthony gave to Prof. Koidl was made available to the public and thereby became state of the art in the sense of Article 54(2) EPC before the relevant priority date 14. September 1989. The same is true for the content of documents D1a and D1b for the reason that they merely exhibit a part of that state of the art.

3. The term "transparent"

3.1. The central plank on which the patentee has chosen to construct its case on novelty is the interpretation that the diamond film is "transparent in the visible spectrum of the light" and that transparency for other parts of electromagnetic spectrum such as infrared and ultraviolet is neither meant nor addressed by the patent. In support of its argument, the patentee has further referred to various dictionaries in order to determine the meaning of "transparent". As opposed to "opaque" and usually distinguished from "translucent", the meaning of "transparent" is explained as having the property of transmitting rays of light so that bodies situated beyond or behind can be distinctly seen (cf. e.g. The Random House College Dictionary, 1982, page 1397).

3.2. Such everyday definitions of transparency are, however, not appropriate to enable in patent claims a clear and unambiguous distinction between the claimed diamond films and those of the prior art. This is even more valid in the present case, since the term "transparency" turns out to be the fundamental key feature. According to the prior art, free-standing diamond films are described as being opaque, brown, greyish, translucent, semi-transparent, or like frosted glass which changes after polishing into "clear". Diamond films can also be in part transparent, namely in the center, whereas the outer ends are opaque. Thus, all grades of transition between the extremes "opaque" on the one hand and "glass-clear" on the other hand do exist. Moreover in its technical definition, the term "transparency" is not confined to the visible part of the electromagnetic wave spectrum, as alleged by the patentee, but also encompasses at least its infrared and ultaviolet regions.

3.3. In the absence of any discrete physical parameter or of a method which allows to define the "transparency" of a diamond film more precisely and in an objective manner, it is impossible to define clear limits for a transparent diamond film in the sense of the claims and for a film which is not "transparent". Consequently transparency is not suitable to distinguish a diamond film falling within the scope of claim 1 from free-standing diamond films of the state of the art, in particular from those forwarded to Prof. Koidl/Fraunhofer by Mr Anthony and/or described in documents D1a and D1b.

4. Analysis and reproducibility of diamond films

4.1. The first condition set out in G 1/92 (cf. point 2.1 above) being met, it has to be considered whether the diamond films could be (i) analysed and (ii) reproduced by the skilled practitioner before the priority date of the patent.

4.2. As to question one, there can be no doubt that in 1988 a skilled expert was able to analyse the physical properties of free-standing polycrystalline diamond films: According to documents D1a and D1b a 400 µm thick sample prepared by hot filament microwave plasma assisted CVD (labelled GE) was investigated at Fraunhofer. The absorption spectrum showed a two-phonon absorption band of diamond together with C-H stretching bands, and an average hydrogen concentration of below 1 at% (= <10000 ppm) is inferred. X-ray measurements to examine the distribution of the crystal orientation revealed a pronounced fibre texture with columnar structures with a preferential alignment of the {110} planes perpendicular to the growth direction (cf. D1a, page 283, page 289, 2nd paragraph, last sentence; page 289, last paragraph, line 4 to page 290, first line; Figures 1, 2, 6, 7). These physical parameters of the sample labelled GE in D1a are identical with those in claim 1 of the patent in suit.

4.3. As to the second question, namely of whether the claimed diamond film was reproducible before the priority date of the patent, the patentee contended for the first time at the oral proceedings before the Opposition Division that the microwave plasma assisted CVD method set out on page 284 of document D1a was not the same as the one applied in the patent in suit and did not lead to transparent diamond films as claimed (see minutes of the oral proceedings before the Opposition Division on 11 May 1998, page 7) 5th full paragraph). In order to disprove this allegation, the opponents OI and OIII submitted comparative tests (documents D15 and D20) in the appeal proceedings. In particular the process parameters selected for performing the experiments HF-MW-1 and 2 described on page 2 of document D15 strictly adhere to those disclosed in document D1a: 1.5 vol% CH4 in H2, 13 mbar = 9.75. torr; tungsten filament temperature 1950 ± 50 C; temperature of the substrate 890 ± 100 C; microwave power 780/750 W, 2.45 GHz; distance between substrate and filament 3 to 8 mm. After removal of the Si-substrate by etching, a 55 µm or 150 µm thick free-standing diamond film was obtained. Due to the polycrystalline structure the diamond films exhibited a rough surface structure but were "transparent to visible light" since a written text could be read through the films (see D15, page 3, paragraph 5.1, Figure 3). The physical parameters obtained from these films comply with those set out in claim 1 of the patent. Based on this evidence it has to be concluded that free-standing polycrystalline diamond films showing a thickness of at least 50 µm and a "transparency" within the meaning encircled in point 3 above could be produced by an expert in the field of CVD.

4.4. The patentee's argument that according to D15 a modern microwave apparatus was used rather than a conventional device that was available in 1988 is not convincing, because the essential parameters, i.e. the frequency (2.45 GHz) and the power of 750 to 780 W, are the same as in the method according to document D1a. Also the free selection of a distance of 3 to 8 mm between the substrate and the hot tungsten filament, which is not expressly indicated D1a and for this reason was objected to by the patentee as ex-post facto knowledge, has no bearing on the matter, because the selected distances are common in this field (cf. e.g. D2, page 4070, lines 1/2; D6, page 451, left hand column, lines 13/14 from the bottom; D10 second paragraph; D19 (translation), page 2, second paragraph, line 4). The same is true in respect of the selection of the substrate which can be silicon or molybdenum. As to the product defined in claim 1 it is irrelevant whether the free standing diamond film is obtained by separation from the substrate (Mo) by cooling or later by etching away the (Si) substrate.

4.5. In view of these considerations, the subject matter of claim 1 lacks novelty with respect to the teaching given in document D1a. Consequently, there is no need to deal with the comparative tests submitted by opponent III or any further state of the art.

5. Reimbursement of fees

The question of reproducibility of the claimed diamond films with the method disclosed in D1a and D1b arose for the first time during the oral proceedings before the opposition division, when Mr Anthony, co-author of D1a and D1b, stated that D1b gives an incorrect method (paragraph bridging page 4 and 5 of the minutes) and the method presented in D1b a and D1b is not the GE method (page 7, on top, of the minutes); however, he refused, as he had been instructed by the proprietor, to divulge the method by which the samples forwarded to Prof. Koidl were made, or to say anything about the alleged incorrectness in D1b (top of page 5, bottom of page 6 and fifth full paragraph on page 7 of the minutes). Opponent III reacted to these submissions by referring to Article 113 EPC and asking for an adjournment so that further tests could be made, because he understood that the proprietor's argument which was being heard for the first time is, that the CVD method (presented in D1a and D1b) does not work, in particular leads to non-transparency (page 6, fourth paragraph of the minutes). The proprietor then explicitly contested that there was an enabling prior art disclosure with a hot filament CVD method (page 7, second full paragraph, of the minutes).

In point 3 of the reasons for the decision under appeal, which was given orally at the end of the oral proceedings, it is set out under the Heading "Novelty" that none of the prior art documents disclosed experiments in which "substantially transparent" diamond films were manufactured and that, in view of the conflicting arguments submitted by the parties, it could not be concluded beyond reasonable doubt that the specimens of diamond films provided by the Proprietor (GE/Mr Anthony) to a member of the public (Fraunhofer/Prof. Koidl) could be manufactured by a skilled person, referring to an enabling disclosure. Therefore, the Opposition Division did not consider these samples as an anticipation of diamond films defined in claim 1 of the patent in suit and, since it took the same view with respect to the other state of the art considered, in particular D1b, eventually it held that the claimed subject-matter was novel.

Having in mind the reasons of the Opposition Division and the considerations of the present decision (see in particular points 3 and 4 above), the decision under appeal probably would have been different, namely revoking the patent in suit, if the opponents had been given an opportunity to refute the proprietor's (Mr Anthony's) late contention, that D1a and D1b were not enabling. Until this contention and the related submissions of the proprietor were made during the oral proceedings, the opponents had, in view of Mr Anthony's behaviour vis-à-vis Fraunhofer/Prof. Koidl and his co-authorship of D1a and D1b, objectively no reason to assume that the enabling nature of the production method described in those documents would be put into question. Rather they, as well as the Opposition Division, were taken by surprise by those new contentions, which not only constitute agere contra factum proprium but are also reproachable because they were brought forward only during the oral proceedings which - for good reasons and as the proprietor must know - aim at a discussion of all relevant aspects of the case so that the decision can be given at the conclusion. However, this is matter of procedural economy which in case of conflict with procedural principles and rights may not override the latter. In the circumstances of the present case, it is the proprietor who is to be blamed for having acted so late, and not the opponents for not having made the tests which were necessary and suitable for refuting the late and portentous contention of the proprietor (cf. experiments shown in D 15 and D 20 submitted during the appeal proceedings). Thus, the refusal to allow such experiments before the decision, which held in favour of the proprietor in respect of the fact in question, was taken, was contrary to Article 113(1) EPC; it was also inappropriate, in that it punished the wrong side by forcing the opponents to appeal against the decision in order to obtain the opportunity to test and to prove the enabling nature of the specific prior art, which is a most relevant issue, as both the decision und appeal and the present decision show. If that had been done before the decision under appeal was taken or if the proprietor's (Mr Anthony's) late contention had been rejected as abusive and/or made in bad faith (see e.g. decisions T 101/97 and T 951/91, the duties of the parties to act at the earliest possible juncture as set out therein being equally applicable to the proprietor), then the patent would already have been revoked by the first instance and the opponents could have saved the fee for appeal at the very least. As regards the costs of possible further oral proceedings before the Opposition Division an equitable apportionment of the costs to the debit of the responsible party (and not of the Office) could then have been achieved on the basis of Article 104(1) EPC.

For these reasons and in view of the outcome of the present appeal proceedings, the reimbursement of the appeal fee to the appellants is both legally possible and equitable by reason of a substantial procedural violation (Rule 67 EPC).