A recent decision by the German Federal Court of Justice (Polyesterabmischungen, X ZR 90/11) relating to a patent concerning polyester resin blends may be instructive to demonstrate how an experimental report reproducing a prior art test instruction can be helpful to a nullity plaintiff, even if the description in the prior art document is incomplete and decisions by the experimenter have to be made when reproducing the prior art teaching.
In practice, the problem sometimes arises that you are supposed to attack a patent with claims that are characterized, inter alia, by a parameter which is not to be found in this form in the prior art. In the case decided, the patent claim read as follows:
9. A polymeric material comprising an aromatic polyester resin and a polyamide derived from dicarboxylic acid and an arilene diamine (sic!), wherein the polyamide is dispersed in a polyester resin matrix with average size of the dispersed domains is from 0.2 to less than 1 micron.
It is hardly surprising that there were already polymer mixtures comprising an aromatic polyester resin and a polyamide made from and arylene diamine and dicarboxylic acid, and also such where the polyamide was dispersed in the polyester resin. Thus, only the last feature was disputed, both with regard to its interpretation as well as its realization in the prior art. The relevant citation was K10. In this respect, the Federal Court of Justice held the following (marginal nos. 28 and 36, underlining added):
2. The Federal Patent Court essentially correctly decided that the subject matter of the version of patent claim 9 defended with the main request is not patentable. It can remain undecided whether it is anticipated by K10 – as assumed by the Federal Patent Court. The reason for this is that it is in any case the result of an obvious implementation of the instruction given to the person skilled in the art in K10, and is therefore not based on inventive step (cf. Federal Court of Justice, Judgment of 24 July 2012 “Leflunomid”, court docket: X ZR 126/09, as published in GRUR 2012, at 1130, marginal no. 27).
Moreover, the Intervenor who has sided the Plaintiff has shown with the test report it submitted in the first instance (as Exhibit SH7) regarding the reproduction of the embodiment example of Table 3 of K10, as carried out by the Fraunhofer Institute for Applied Polymer Research, that a composition can be obtained in this manner, in which firstly the arithmetic mean of the domain sizes 0.2405 µm and the median 0.2319 µm are close to each other since there is almost a normal distribution of the domain sizes, and secondly both values are higher than 0.2 µm and thus lie within the range claimed with patent claim 9. Defendant’s objection, i.e. that the mixture on which the test report SH7 is based does not represent a reproduction of K10 since neither the mixing conditions nor the polyamide component are given in K10 and they are thus randomly selected, is not convincing. According to the list of the components provided in the test report, the composition examined therein comprises all components of the composition suggested in claim 10 of K10. Since claim 10 of K10 does not provide any statements regarding the polyester to be used and the transition metal catalyst, and only stipulates with regard to the polyamide that it has to contain a xylylene group, the mixture chosen for the test according to SH7 is also comprised by K10. Thus, the person skilled in the art can obtain the domain size set forth by Plaintiff and its Intervenor by implementing the frame provided by K10 lege artis. Defendant has not asserted that this was not the case in the test by the Fraunhofer Institute. The court expert also did not see any indications for this.
Put in simple words, this means that an experimental reproduction of a prior art teaching, that is carried out lege artis, is sufficient to at least affirm obviousness, even if the reproduced instruction has some gaps which need to be filled (as so often in real life).