Abrasive outcome on over-broad ranges for 3M in Saint-Gobain dispute re-emphasizes importance of narrow dependent claims

Any litigator knows that sometimes life is not fair.  With all the best practice, best arguments and right on your side, you still fall down.  The Patent Gods aren not always benevolent forces, however.  This seemed to be the story of 3M in its litigation with Saint-Gobain who defeated so many attacks, but to fell down on the final attack.  The case raises the importance for patent attorneys and litigators alike - it is good to be armed with a suite of narrower dependent claims in prosecution or conditional amendment in litigation.  Our Kat Friend in Brussels Henry Yang investigates.  

Over to Henry:

"One could not help but feel sorry for the European Patent (UK) 2 373 755 (‘the Patent’), owned by company 3M. Having resisted so well to the inventive step attacks and held its ground against the arguments on undue burden to produce anything within the claims, it was ultimately felled by Deputy Judge Michael Tappin QC in Saint-Gobain v 3M [2022] EWHC 1018 back in May because there was no dependent claim with a narrower range to save the independent claim whose range was too broad.  So for a patent all about abrasion, one can imagine how abrasive this outcome was for 3M.  Here is what happened.  

The Patent

The Patent is entitled ‘Dish-shaped abrasive particles with a recessed surface’. There are two main embodiments in the Patent. The first, shown in Figs. 1A and 1B, are called planar/concave particles in the judgment (paras 62 – 64):

The second main embodiments are illustrated in Figs. 3A and 3B, called concave/convex particles in the judgment (paras 76 – 78):

The particles (the parts without hatching in Fig. 1C) are fixed to the backing 42 to form a coated abrasive article 40 (which may be in the form of a disc, belt or sheet) in a way shown below:

To produce the shaped abrasive particles, the Patent mentions seven process steps: (1) providing a dispersion, (2) providing a mould, (3) filling the mould with the dispersion, (4) controlling the rheology of the sol-gel in the mould to make different types of dish-shaped particles, (5) removing the precursor dish-shaped abrasive particles from the mould, (6) calcining and (7) sintering (para 83). Claim 1 was broken down into integers as follows (para 93):

Interpretation of the claims

Saint-Gobain (SG) tried to sever integer (6) from integer (9). They submitted that it was possible for a collection of particles none of which has the Tc/Ti ratio of between 1.25 and 5.00 to satisfy as a collection the integer (9) test. They said that it was also possible that a collection of particles most of which have that Tc/Ti ratio fails the integer (9) test. But the Judge pointed out that integer (9) provides the test for determining the thickness ratio of integer (6). There is no separate assessment of individual particle Tc/Ti ratios for integer (6) (para 103).

SG also questioned how a skilled person determines whether a particle is dish-shaped (para 107). The Judge held in para 109 that the skilled person would be capable of doing so by referring to the teaching of the Patent, which (i) identifies exemplary dish-shaped particles, (ii) distinguishes them from the essentially flat particles in a prior art document US 5 366 523 (Rowenhorst), and (iii) explains that these dish-shaped particles have grinding benefits compared to the essentially flat particles because of their dished shape.

Validity over US 5 366 523 (Rowenhorst)

Rowenhorst discloses amongst others essentially flat particles, compared with the particles in the Patent (para 69). An example is given in Fig. 5 of the Patent:

SG’s arguments had three facets. First, it was inevitable or would happen frequently that carrying out example 1 or 9 of Rowenhorst would produce particles within the claims. Secondly, the particles within the claims could come from obvious modifications to Rowenhorst. Thirdly, the particles of the claims made no technical contribution to Rowenhorst (para 129).

On the first point, the Judge noted that the evidence did not establish that carrying out example 1 or 9 of Rowenhorst would produce batches with particles with individual Tc/Ti ratios at 1.25 or greater (para 132). In addition, SG only provided the Tc/Ti measurements of two Rowenhorst particles. But these particles were not representatives of the 5% of the sample which had some degree of curvature (para 133).

On the second point the Judge started by considering whether the skilled person would have any motivation to conduct workshop modifications to Rowenhorst (para 143). Specifically, the case focused on the difference between example 9 of Rowenhorst and example 1 of EP 2 373 458 (para 146) (the Patent referred to the US patent application 12/337,001; EP 2 373 458 is derived from US ’001 and contains the same relevant disclosure as US ’001). SG’s expert said that by changing the variables in Rowenhorst there was a chance that one could hit the conditions leading to example 1 of the Patent. He also mentioned that lots of different outcomes could be obtained just by changing the parameters in Rowenhorst (para 147). The Judge pointed out that this was a typical example of a step-by-step analysis infected by hindsight, for it relied on the skilled person trying a combination of conditions when several parameters could be changed (para 148). Another reason why this point failed was that Rowenhorst emphasised ensuring that the end abrasive particles are flat. Even if the skilled person were to produce by mistake precursor particles which are not flat but planar/concave, according to Rowenhorst he would discard them instead of continuing the steps to obtain the end abrasive particles (para 150).

On the third point, the technical effect of the Patent is that dish-shaped particles grind better compared to the essentially flat particles of Rowenhorst. The Patent presents two theories: the ice cream scoop theory – an ice cream scoop with a concave shaped end effectively digs into materials and removes a significant quantity of them, ("Yes, please", says Merpel) and the thinness theory – the thin interior of the abrasive particles not only means that the flat worn down area is smaller, but also increases fracturing of the particles during use which resharpens the particles (paras 156; 59; 65). On the thinness theory, the Judge noted that it depends on the absolute thickness of the interior rather than on the shape or thickness ratio of particles. This theory is therefore not capable of supporting the claimed invention across its breadth (para 158). On the ice cream scoop theory SG submitted that any otherwise plausible theory was disproven by the data (para 161), but the Judge did not read that the data in the Patent do so (para 164). This meant that the Patent does disclose a plausible technical effect through the ice cream scoop theory associated with the claimed invention (para 168).

Insufficiency

Uncertainty-type insufficiency

SG’s arguments on uncertainty-type insufficiency had two points. First, the claim did not specify how to decide if a particle is dish-shaped and/or could be included in the integer (9) test. This has already been addressed in the claim interpretation above. On whether a particle is dish-shaped, the Judge indicated that both Kirin-Amgen [2004] UKHL 46 and Anan Kasei [2019] EWCA Civ 1646 distinguished a fuzzy boundary from conceptual uncertainty. He noted that while here it was a question of degree which would give rise to a fuzzy boundary, this did not mean insufficiency (para 180).

The second point was that the results of the integer (9) test could differ depending on the selection of dish-shaped particles (para 181). The Judge held that a mere theoretical possibility that different selections would give materially different average Tc/Ti ratios was not enough. It was necessary that the uncertainty in the test could give rise to a difference in the result obtained in practice (para 184). The Judge also mentioned that a part of this assessment involved considering whether there are other sources of experimental error which may overwhelm the variation in result which is due to different versions of the test (para 184). There was no evidence addressing these questions, and therefore SG did not succeed in showing that the integer (9) test, which allows for different random selections of particles, gives rise to a practical problem in performing the invention (para 190).

Undue burden to produce anything within the claims

SG made two points on undue burden. First, there is an undue burden to produce anything within the claims. Secondly, there is an undue burden to produce particles with an average Tc/Ti ratio at the upper end of the claimed range (para 198).

On the first point SG’s expert indicated that in example 1 of the Patent there are several uncertainties, including the rheology of the dispersion; the level of adhesion of the dispersion to the mould; and the drying rate (para 202). But the Judge considered that the evidence did not establish when a certain rheology or mould material (related to the level of adhesion) or the airflow level (related to the drying rate) would lead to failure to produce planar/concave particles with the level of dishing indicated in example 1 while another would lead to success (paras 204 – 206). The problem, as the Judge explained in para 209, is that the range of values for the relevant parameters which would lead to planar/concave particles with the degree of dishing of example 1 was unknown. It was therefore impossible to know how likely the skilled person, when implementing example 1, would succeed at the first attempt or find a successful set of conditions by routine trial and error. SG has therefore failed to prove that it would require undue burden to make any particles within the claims (para 209).

Undue burden to produce particles across the scope of the claims

The Patent mentions that triangular dish-shaped particles produced according to the invention have Tc/Ti ratios between 1.55 to 2.32 in some embodiments. There is no indication in the Patent that it is possible to obtain particles with those ratios above 2.32, in particular towards the upper end of the claimed range at 5.00 (para 211). 3M’s expert proposed a drying theory where he said that the faster the drying rate the more pronounced the warping: to increase warping the skilled person would consider increasing the drying rate (para 213), but he also accepted that it was difficult to predict the effect of increasing the temperature. He also agreed that increasing the temperature may have side effects: increase the risk of fracture of the particles and of degradation of the mould. The Judge decided based on the evidence that the skilled person would not come up with the drying theory to increase the Tc/Ti ratio (para 217). 3M’s expert also proposed an inverted moulding theory where he suggested that a concave lower face of the particles could be formed by controlling the degree of curvature of the upper face through how the particles are dried (para 215). But he also agreed that the Patent contains no guidance on the dispersion or drying conditions or the amount of release agent for this, and that the final shape of the particles would be very difficult to predict. The Judge concluded that the skilled person would not come up with the inverted moulding theory to try to increase the Tc/Ti ratio either (para 218).

Finally, 3M tried to rely on the cross-reference to information in US ’001 (which as noted above contains the same relevant disclosure as EP ’458), but the Judge would need expert evidence to see if the skilled person could, based on the statements in EP ’458 and without undue burden, to produce particles with an average Tc/Ti ratio near the upper end of the claimed range by increasing the drying rates while using a mould release agent (para 228). The Judge also wondered whether the skilled person would consult US ’001 at all: he considered that for the skilled person to consult a document for a particular purpose, Halliburton v Smith [2005] EWHC 1623 (Pat) requires that it must be clear that the skilled person should consult the document for that purpose. 

He also considered that ‘the question of what a skilled person would do, when faced with a cross-reference and the document cross-referred to, is one on which expert evidence is admissible, and indeed required if a party wishes to establish that the skilled person would alight on particular text and find it of assistance in a particular way’ (para 229). Here, there is no indication in the Patent that US ’001 has information to make dish-shaped particles with Tc/Ti ratios towards the upper end of the claimed range. In addition, evidence is required that the skilled person would refer to the passages relied on by 3M and understand that they provide a solution to making dish-shaped particles with Tc/Ti ratios towards the upper end of the claimed range (para 230), but no such evidence was provided (para 231).

Consequently the skilled person could not make dish-shaped abrasive particles with Tc/Ti ratios towards the upper end of the claimed range: claim 1 is invalid due to insufficiency. As no dependent claim contains a limitation to that claimed range the Patent was revoked (para 233).

Comment

From the discussion on undue burden to produce anything within the claims in this judgment (paras 199 to 210) it shows that the existence of a mere burden to produce anything within the claims is not enough to prove insufficiency: demonstrating how undue this burden is is more crucial. It would be preferable to show the range of parameters that can lead to the claimed product, and at the same time how easy it is for the skilled person to step outside that range. As the Judge vividly put it, there is a difference between ‘scattered islands in a sea of failure’ when ‘finding those islands without navigational aids’ is unduly burdensome, and ‘sunlit uplands with scattered sinkholes’ when it is not.

Still regarding insufficiency, para 229 of this judgment contains a short discussion on when information in a cross-referred document can be used. The Judge’s interpretation of Halliburton v Smith may be strict, but in this guest Kat's opinion, it is quite appropriate to prevent the skilled person from being overwhelmed by irrelevant information in a referred document: the phrase ‘the content of application A is incorporated in application B in its entirety’ is a common sight in many applications, and letting this general statement to fill in specific gaps in an application looks too generous to the applicant.

Finally, a natural lesson from this case is the importance of providing ample narrower ranges in the dependent claims which correspond better to the experimental results, if the applicant broadened them in the independent claims and successfully obtained wide protection: the Patent could have been saved by a claim limiting the Tc/Ti ratio to 1.55 – 2.32."