Sandvik v Kennametal: the dangers of armchair examples

An important part of getting a valid patent is that it needs to disclose the invention in such a way that it can be reproduced without undue effort by a nominal 'skilled person'. This requirement is enshrined in Article 83 EPC, which simply states:

The European patent application shall disclose the invention in a manner sufficiently clear and complete for it to be carried out by a person skilled in the art.

This requirement is normally not something that can be tested during examination. It is often therefore just taken as read that whatever is described in a patent application could in fact be carried out by a skilled person simply following one or more of the examples in the description. There should, of course, be at least one example in the description that results in a product, or follows a process, that lies within the scope of the claimed invention.

In some cases the applicant may be so confident the invention will work as planned that examples are provided setting out how to make a product according to the invention without this having in practice being done when the application is filed.  This is normally nothing to be too worried about, but the applicant should always be aware of the potential fatal flaw in the application if the example does not in fact work.

Such a fatal flaw was revealed in a recent judgment by Mr Justice Arnold in the case of Sandvik v Kennametal, available from BAILII here. Sandvik had obtained a European patent, EP0603144, relating to an alumina-coated cutting tool (similar to those shown in the picture on the right). The key feature of the claimed invention was that the alumina coating had a particular texture, i.e. a preferred crystallographic orientation relative to the cutting face.  The claims specified a 'texture coefficient' (TC) of larger than 1.3 in relation to one particular orientation, which meant that this orientation would be preferred over the others. The claims also specified how the texture coefficient should be measured, which was by using standard X-ray diffraction data patterns.

Unfortunately for Sandvik, when an experiment was carried out following one of the 'armchair examples' (as Arnold J put it: paragraph 147), the resulting texture coefficient was much less than expected. The best that could be done was a TC of 0.84, which related to the particular orientation being less preferred rather than more preferred. Sandvik tried to argue that the skilled person would be able to get to the required result specified in the claims by doing a few trial runs, as they had done when optimising the process. However, the key feature that allowed the claimed texture coefficient to be obtained, relating to the supply of gases during deposition of the coating, was not disclosed in the patent and was even admitted by Sandvik not to have been discovered until after the priority date of the patent.

Arnold J considered that this missing information was fatal to the patent.  Without the skilled person being informed of how to get the required texture, it was not possible for the claimed invention to be sufficiently disclosed because the information on what to do and why was not part of the common general knowledge. As Arnold J put it, when following the example outlined in the patent "the skilled reader would not appreciate that he was doing something that he was not supposed to do" (paragraph 137). The patent was therefore found to be invalid on the ground of insufficiency. It was also found invalid for being obvious, but this was something of a moot point (unless the insufficiency finding is overturned on appeal).

The main lesson to be learned from this case is fairly obvious, which is that examples of how to perform an invention should be sufficiently described in enough detail for someone else to be able to do it. It is dangerous to assume that the skilled person can fill in any gaps, as these could be too wide when it comes down to proving it.