Wooly invalid patent

So, this is probably not a very good pun to start a new year with; but you have to admit that there are no easy puns with mineral wool, which was the subject-matter of today’s patent – a patent which was found to lack inventive step by Board of appeal 3.3.05.

The problem and solution approach is a must for any inventive step reasoning at the EPO. The first step of this approach consists in determining the closest prior art.

However, practically speaking, it is also very useful to run the approach backwards: if, as the patentee, you realize that, starting from prior art document A, it is much easier to arrive at a conclusion of non-obviousness than starting from prior art document B, then you know you need to push for the selection of A as the closest prior art, and thus come up with arguments disqualifying B. Of course, the opposite is true if you are an opponent.

One strategy sometimes used by patentees to get around a prior art reference which would make their lives miserable if selected as the closest prior art is to explain that this prior art reference is not a realistic starting point for the skilled person.

A typical example of a document which would not be a realistic starting point is one which has a lot of features in common with the claimed invention, but in a different context. As stated e.g. in T 870/96, the closest prior art should be a document

which [the] skilled person would have realistically taken under the “circumstances” of the claimed invention insofar as these circumstances can be retrieved in one item of the prior art. Consequently, among these “circumstances”, aspects as the designation of the subject matter of the invention, the formulation of the original problem and the intended use and the effects to be obtained should generally be given more weight than the maximum number of identical technical features (reasons, 4.1).

Now, in today’s decision T 2579/11, inventive step of the opposed patent had to be assessed in view in particular of two relatively similar prior art references, documents A5 and A6. These prior art references were in fact similar for a good reason: A5 was the priority document of patent application A6.

The patentee argued that the closest prior art was A6 and not the priority document A5. It is straightforward to guess why the patentee made such a submission: the claimed invention related to a mineral wool comprising a number of compounds within specific concentration ranges or concentration ratios. Both A5 and A6 disclosed mineral wools containing the same compounds as the claimed invention. But the examples disclosed in A6 differed from those of A5. In examples 2 and 5 of A5, only the concentration ratio MgO/CaO was outside of the claimed range. In contrast, in some of the examples of A6, not only was the MgO/CaO ratio outside of the claimed range, but in addition the alumine content was above the claimed maximum threshold. Or, in other examples, the alumine content was within the claimed range but then the MgO/CaO ratio was further away from the claimed values than in examples 2 and 5 of A5.

As a result, it was easier to arrive at a finding of lack of inventive step starting from A5 than starting from A6, since fewer modifications needed to be made starting from A5.

Artificial sheep from which mineral wool can be sheared
Artificial sheep from which mineral wool can be sheared

In order to try to disqualify A5 as the closest prior art, the patentee offered the following argumentation:

A5 is accessible only in the examination file wrapper of A6. The skilled person has no reason to start from priority document A5 since he/she has a more elaborate text available, which was filed after one year of additional research. The fact that examples 2 and 5 of A5 were not included in A6 shows that they were not valid examples. Therefore, document A6 is the closest prior art (facts and submissions, VII).

But this did not fly with the Board. The Board first of all noted that both A5 and A6 had the same purpose (which was also the purpose of the claimed invention), namely providing mineral wool which can be dissolved in a physiological medium, and which can be fibered by centrifugation (reasons, 2.2.5). Second of all, in terms of features in common with the claimed invention, it was concluded that, for the reasons already set forth above, “the wools according to examples 2 and 5 of A5 request fewer structural modifications than the wools of A6” (reasons, 2.2.6).

Finally, the Board addressed the patent proprietor’s argumentation as follows:

Once a priority document has been made accessible to the public, it belongs to the state of the art under article 54(1)(2) EPC like any other publicly available document. The papers of a patent application on the one hand and a priority document on the other hand are then entirely distinct documents, even if the latter is made accessible to the public through the publication of this application. The fact that a later application generally has a more detailed description than the priority application can therefore not, by itself, justify that the priority document should be discarded in favor of the later application when the closest prior art is selected. The fact that an embodiment and/or an example contained in the priority application are not included in the application claiming the priority does not by itself allow one to conclude that such embodiment and/or example were less preferred, or less valid. 

[…] Even assuming that the later application A6 contains results of complementary research, nothing indicates that examples 2 and 5 of A5 which are not included in A6 would be invalid or of poor quality or insufficiently disclosed (reasons, 2.2.7).

Thus, examples 2 and 5 of A5 were deemed to represent the closest prior art, and the rest of the problem and solution approach analysis led the Board to the revocation of the patent due to lack of inventive step.

To some extent, it seems to me that a parallel could be made with other cases such as T 741/91 or  T 334/92, where an old prior art document was rejected as a candidate closest prior art because said document had been ignored in the relevant field and had not given rise to further development.

Here, the priority document A5 was dated at most one year prior to the subsequent application A6, so this was not a very”old” document by any means. But the common trait is that, in both situations, an argument was made that a prior art reference should be disregarded as the closest prior art, not because of its intrinsic contents, but rather because of extrinsic evidence that the reference would not count in the eyes of the skilled person. In T 2579/11 the argument failed, whereas it succeeded in T 741/91 or T 334/92. This may just be a matter of the extrinsic evidence being persuasive enough or not.

As a side remark on a procedural aspect of the present decision, the Board skipped the novelty objections and only took position on inventive step as this was sufficient to reach the conclusion that the patent should be revoked. Similarly, the Board focused on inventive step of the most restricted claims defended by the patent proprietor, namely auxiliary request No.3, the conclusion on this request also applying to the previous, broader requests.

This is a reminder that Boards of appeal have a lot of leeway for dealing with submissions of the parties in the order that they see fit, and for getting to the final decision in the most efficient manner. However, shortcuts such as the present one may seem somewhat unorthodox. In particular, there was an objection of lack of novelty of the main request and the first auxiliary request over A5 (as well as over another document A3). This means that the Board decided that these requests involved no inventive step over A5 without ruling whether the claims at stake were even novel over A5.


CASE REFERENCE: Board of Appeal 3.3.05, T 2579/11, Laine minérale / Saint-Gobain, March 10, 2015