Docket: A-379-14
Citation:
2016 FCA 216
|
CORAM:
|
WEBB J.A.
BOIVIN J.A.
DE MONTIGNY J.A.
|
|
BETWEEN:
|
|
NOVA CHEMICALS
CORPORATION
|
|
Appellant
|
|
and
|
|
THE DOW
CHEMICAL COMPANY,
DOW GLOBAL
TECHNOLOGIES INC. AND DOW CHEMICAL CANADA ULC
|
|
Respondents
|
REASONS
FOR JUDGMENT
DE MONTIGNY J.A.
[1]
This is an appeal from a decision of Justice
O’Keefe (the Judge) granting an action for patent infringement brought by the
Dow Chemical Company, its subsidiary Dow Global Technologies Inc. and its
licensee Dow Chemical Canada ULC (collectively Dow, or the respondents) against
the Nova Chemicals Corporation (Nova or the appellant) in respect of its
product SURPASS. The parties are both manufacturers of polyethylene film-grade
copolymers for use in, amongst other things, packaging applications. The Dow
Chemical Company is the owner of Canadian Patent 2,160,705 (the ‘705 Patent)
for “Fabricated Articles Made from Ethylene Polymer
Blends”.
[2]
The Judge confirmed the validity of the ‘705
Patent, which was challenged on the grounds of utility, over-claiming,
anticipation, obviousness, double-patenting and insufficiency of the
specification. The Judge also construed six disputed terms in the patent
claims, which led him to conclude that Nova had infringed the patent. In coming
to his conclusions, the Judge largely preferred Dow’s experts and fact
witnesses over Nova’s own expert witnesses.
[3]
Before this Court, Nova appeals all of the Judge’s
findings regarding validity and also contests the construction of four terms in
the ‘705 Patent, as well as the corresponding infringement findings. Having
carefully considered the oral and written arguments of the parties, I am of the
view that the appeal ought to be dismissed; the appellant has failed to
demonstrate any error of law, and it is not the role of this Court to
second-guess the Judge’s assessment of the evidence that was put to him.
I.
The Patent at Issue
[4]
The ‘705 Patent was filed in Canada on April 19,
1994. It claims priority from US patent application US08/054,379 dated April
28, 1993. It was published on November 10, 1994, and expired on April 19, 2014.
[5]
The patent is directed primarily to polyethylene
used to make “film” products, i.e. sheets of
plastic, like plastic garbage bags and food wrapping. Some film applications do
not have demanding strength requirements, but others do. One solution for these
demanding applications was to make thicker “films”
so that they are stronger. That requires the use of more plastic, however,
leading to higher costs and more waste when the plastic film is disposed of.
[6]
The patent identifies the need to develop
polymers that can be formed into thinner films with improved strength
properties:
There is a continuing need to develop
polymers which can be formed into fabricated articles having these combinations
of properties (e.g., improved modulus, yield strength, impact strength and tear
strength, preferably greater dart impact for a given yield strength in the c[a]se
of films and greater IZOD impact for molded parts). The need is especially
great for polymers which can be made into film which can also be down gauged
without loss of strength properties, resulting in savings for film
manufacturers and consumers, as well as protecting the environment by source
reduction.
‘705 Patent, p. 1, lines 22-31
[7]
The person skilled in the art (POSITA) would
have appreciated that there was often a trade-off between film properties such
as yield strength, impact strength, and toughness. The POSITA would also have
known that there is a trade-off with polymer materials, including film, between
strength and toughness. Efforts to modify a polymer material to improve
strength often resulted in reduction of toughness, as mentioned in the ‘705
Patent:
Previous attempts were made to optimize film
tensile strength and yield strength by blending various heterogeneous polymers
together on theoretical basis. While such blends exhibited a synergistic
response to increase the film yield strength, the film impact strength followed
the rule of mixing, often resulting in a “destructive synergism” (i.e., the
film impact strength was actually lower than film made from one of the two
components used to make the blend).
‘705 Patent, p. 1, lines 10-17
[8]
The claimed invention and Dow’s commercial
embodiment of it (ELITE) allows for source reduction to make thicker films
thinner, but just as strong. Whereas prior art efforts to create improved
polymers and polymer blends were largely trial and error, Dr. Lai (one of the
inventors) testified at trial that Dow’s researchers took a different approach
to identify the optimal blend based on polymer density, molecular weight, and
strain hardening (the latter being a property wherein a material becomes harder
as it is stretched). This work is disclosed in the ‘705 Patent, including the
creation of the slope of strain hardening coefficient (SHC) to identify
polymers of interest. The ‘705 Patent describes the invention in the following
terms:
Surprisingly, we have now discovered
compositions useful in films and molded parts having synergistically enhanced
physical properties, which compositions comprise a blend of at least one
homogeneously branched ethylene/α-olefin interpolymer and at least one
heterogeneously branched ethylene/α-olefin interpolymer.
In particular, formulated
ethylene/α-olefin compositions have now been discovered to have improved
physical and mechanical strength and are useful in making fabricated articles.
Films made from these novel compositions exhibit surprisingly good impact and
tensile properties, and an especially good combination of modulus yield,
ultimate tensile, and toughness (e.g. dart impact).
‘705 Patent, p.1, line 32 to p. 2, line 6
[9]
Each of the 46 claims of the ‘705 Patent is
directed to a blend having at least these two components, with each component
having certain requirements, depending on the particular claim. While the
claims recite several limitations, the major disputes in the case revolve
around just two of them. For Component A, the dispute is over the claim term “slope of strain hardening coefficient”. For Component
B, the main dispute is over the claim term “heterogeneously
branched”. All the asserted claims include these terms.
[10]
Dow filed a Statement of Claim on December 9,
2010, alleging that Nova was infringing their ‘705 Patent. Nova counterclaimed
on the grounds of invalidity and unjust enrichment, but eventually dropped its
unjust enrichment claims. In its opening statement at trial, Dow restricted the
litigation to only eight composition claims, being claims 11, 29, 30, 33, 35,
36, 41 and 42; Nova similarly restricted its invalidity counterclaim to these
same claims. As a result, the Judge erred in holding that Claim 15 was valid
and infringed; Dow dropped its allegations in relation to that claim, and
reference to it in paragraph 1 of the Judgment should be deleted.
[11]
The trial began on September 9, 2013 and lasted
for 32 days. Nova called three expert witnesses (Dr. Charles Stanley Speed, Dr.
Francis Mirabella and Dr. Mukerrem Cakmak) and three fact witnesses, while Dow
called three expert witnesses (Dr. Joao Soares, Dr. Robert Young, and Dr.
Christopher Scott) and one fact witness. The Judgment was released on May 7,
2014 and the Reasons for Judgment were issued on September 5, 2014. The Judge
found that all the claims at issue were valid, and that Nova infringed these
claims by manufacturing in Canada and distributing, offering for sale, selling
or otherwise making available film-grade polymers under the name SURPASS.
II.
Issues
[12]
Nova contends that the Judge made a number of
reviewable errors. I believe these issues can be narrowed down as follows:
As for the validity of the ‘705 Patent:
1)
Did the Judge err in deciding that there was no
promise of synergistic utility?
2)
Did the Judge err in finding that the invention
was not obvious?
3)
Did the Judge err in finding that the claims
were not broader than the invention made or disclosed?
As
for infringement:
4)
Did the Judge err in inferring that SHC is
determined using a load/elongation curve?
5)
Did the Judge err in construing “heterogeneously branched”?
6)
Did the Judge err in construing “comprising”?
7)
Did the Judge err in deciding that the high
density (HD) fraction was at least 5% by weight of the composition?
III.
Analysis
[13]
There is no dispute between the parties as to
the applicable standard of review. Pursuant to the decision of the Supreme
Court in Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235, the
standard of review for findings of fact is palpable and overriding error, and
correctness for errors of law. The standard is also palpable and overriding
error for findings of mixed fact and law, unless there is an extricable
question of law, which will be reviewed on a correctness standard.
[14]
It is worth emphasizing at this stage that a
trial judge is entitled to deference in its appreciation of the evidence,
particularly the expert evidence that affects the construction of a patent. In
particular, the judge is entitled to weigh the evidence and to prefer certain
evidence. It is certainly not the role of an appeal court to retry the case or
to second-guess the trial judge’s assessment of the factual record and of the
evidence. An appellant will consequently bear a heavy burden on matters of
fact. As this Court held in Canada v. South Yukon Forest Corporation,
2012 FCA 165 at para. 46, 431 N.R. 286, “it is not
enough to pull at leaves and branches and leave the tree standing” when
arguing palpable and overriding error; “[t]he entire
tree must fall”.
[15]
On the other hand, the construction of the
patent is to be reviewed on the basis of correctness. As the Supreme Court has
stated in Whirlpool Corp. v. Camco Inc., 2000 SCC 67 at para. 61, [2000]
2 S.C.R. 1067, “claims construction is a matter of law”.
That being said, I share the concerns of my colleague Justice Stratas that it
will often be difficult, if not unrealistic and artificial, to distinguish
between those aspects of claim construction that flow from the trial judge’s
assessment of expert evidence from the words of the claim themselves (see Cobalt
Pharmaceuticals Company v. Bayer Inc., 2015 FCA 116 at paras. 16-24, [2015]
F.C.J. No. 555). After all, the construction of a patent is heavily dependent
on the evidence given by persons skilled in the art, and that evidence will
bear heavily on the judge’s findings. For that reason, I accept (as I must)
that the construction of a patent is a question of law to be reviewed on a
standard of correctness, but trial judges are nevertheless entitled to some
leeway as they are often in a much better position than appellate judges to
understand the intricacies of the art underlying the invention disclosed in a
patent.
1)
Did the Judge err in deciding there was no promise
of synergistic utility?
[16]
Section 2 of the Patent Act, R.S.C.
1985, c. P-4, defines an invention as “any new and
useful art, process, machine, manufacture or composition of matter, or any new
and useful improvement in any art, process, machine, manufacture or composition
of matter”. As a result, a patent holder must be able to demonstrate the
utility of the invention or to show that it could be soundly predicted at the time
the patent was applied for (see Apotex Inc. v. Wellcome Foundation Ltd.,
2002 SCC 77 at para. 46, [2002] 4 S.C.R. 153, cited by this Court in Apotex
Inc. v. Sanofi Aventis, 2013 FCA 186 at para. 46, [2015] 2 F.C.R. 644 [Apotex
& Sanofi FCA]). An inventor does not need to describe the utility of
his/her invention, in which case a “mere scintilla”
of utility will be sufficient. If, however, an inventor does make an explicit
promise of a specific result, utility will have to meet that promise; it will
be, in the words of the Supreme Court in Free World Trust v. Électro Santé
Inc., 2000 SCC 66 at para. 51, [2000] 2 S.C.R. 1024, a “self-inflicted wound” (see also Apotex &
Sanofi FCA at paras. 54 and 65-67).
[17]
The Judge also clearly set out the process
whereby the existence and scope of any promise is to be determined: 1) one must
first look for the elevated promise or claimed utility in the claims of the
patent; and 2) then consider any statement found elsewhere in the disclosure,
which should be taken as “mere statement of advantage”
unless the inventor “clearly and unequivocally”
states that it is part of the promised utility of the invention (Reasons for Judgment,
para. 183).
[18]
This approach is broadly consistent with the
most recent jurisprudence of this Court and of the Federal Court, and Nova did
not challenge the soundness of the Judge’s reasons in this respect (see Apotex
& Mylan v. Pfizer, 2014 FCA 250 at paras. 65-66 and 76, 465 N.R. 306 [Apotex
& Mylan]; Fournier Pharma Inc. v. Canada, 2012 FC 741 at para.
126, [2012] F.C.J. No. 901; Bauer Hockey Corp. v. Easton Sports Canada Inc.,
2010 FC 361 at para. 290, [2010] F.C.J. No. 431, aff’d 2011 FCA 83, 414 N.R.
69, quoting, at para. 182, H.G. Fox, Canadian Patent Law and Practice, 4th
ed. (Toronto: Carswell, 1969) at pp. 152-154). The Judge’s task was therefore
to construe the patent to determine if a POSITA would understand it to contain
an explicit promise that the invention will achieve a specific result.
[19]
There is some ambiguity as to what Nova
considers to be the promise made by the patent. At trial, Nova relied on one of
its experts, Dr. Speed, whose position was understood by the Judge to be that “all the blends contemplated by the Patent and every property
of all the blends must exhibit a synergistically enhanced physical property”
(see Reasons for Judgment, para. 186). It is true that on cross-examination,
when pushed on that question, Dr. Speed answered that “at
least some of those properties” would show particularly good performance
(Cross-Examination of Dr. Speed, Appeal Book Vol. 19, Tab 355 at p. 18429).
This is not, however, how Dr. Scott understood Dr. Speed’s expert report (see Expert
Report of Dr. Scott, Appeal Book Vol. 35, p. 6033 at para. 78), and the Judge
cannot be faulted for having read Dr. Speed’s report the same way Dr. Scott
did.
[20]
In its written closing submissions, counsel for
Nova summarized the promise of the patent as being that “at least impact strength will be better than the rule of
mixing when other properties like tensile strength, yield and modulus are
surprisingly improved in the newly discovered compositions”. And on
appeal, Nova framed the promise in a slightly different manner, speaking of
compositions having “improved properties”, “impact and tensile properties at a level above the rule of
mixing”, and “synergistically enhanced physical
properties, surprisingly good tensile and impact properties, or any improvement
in the properties” (see Memorandum of Fact and Law of Nova at paras. 28,
40 and 45). Needless to say, such variations in the formulation of the impugned
promise detract from the requirement that the promise be specific and clear.
[21]
Be that as it may, and however the alleged
promise is interpreted, the main problem with Nova’s position is that it rests
exclusively on the use of the words “synergistically
enhanced properties” used on two occasions at page 1 of the patent, previously
quoted in paragraphs 7 and 8 of these reasons.
[22]
After quoting from both Dr. Speed’s and Dr.
Scott’s expert reports, the Judge found that the inventors had not made an
explicit promise of a specific result, and that the ‘705 Patent had rather been
shown to be concerned with particular compositions that were useful. He came to
that conclusion essentially for two reasons. First, he found no reference in
the claims of the patent to an elevated promise of “synergistically
enhanced physical properties”, nor to any specified level of
improvement; in the Judge’s view, the claims simply speak of the compositions
having improved properties. Second, he preferred Dr. Scott’s testimony over
that of Dr. Speed, because the former addressed the patent as a whole, while the
latter ignored the parenthesized words in the patent defining “destructive synergism” as well as the examples given
in the patent.
[23]
Nova challenges those findings on a number of
bases. It asserts that the Judge gave too much weight to the presumption that a
statement of utility found elsewhere than in the claims of the patent was a
mere statement of advantage, and not a promise. I disagree. The Judge did not
rely on such a presumption, and there was indeed no need to rely on that
expediency.
[24]
However one looks at the ‘705 Patent, two
crucial facts stand out: there is not a single statement of utility in Dow’s
claims, as acknowledged by Nova, and there is only one reference elsewhere in
the specification to support an argument of enhanced utility. I agree with Dow
that the Court should be wary of using a stray phrase on page 1 of the patent
to define the promise of the patent. Such an approach would clearly be
inconsistent with another presumption that an inventor should only be held to
an elevated standard where a clear and unambiguous promise has been made. As
this Court stated in Apotex & Mylan at paragraph 66, “[w]here the validity of a patent is challenged on the basis
of an alleged unfulfilled promise, the patent will be construed in favour of
the patentee where it can reasonably be read by the skilled person as excluding
this promise”.
[25]
The Judge did not make a palpable and overriding
error in finding that the first reference to “destructive
synergism” found in the second paragraph of page 1 of the patent does
not relate to the construction put forward by Dr. Speed that the patent
promises an elevated level of utility. This passage is clearly descriptive of
prior attempts at blending various heterogeneous polymers and the trade-off
problem encountered.
[26]
As for the second reference to “destructive synergism” at page 1 of the patent (the
only relevant one for the purpose of identifying an explicit promise of a
specific result), I am also of the view that the Judge could properly find that
it is far from sufficient to infer any specified level of improvement. It does
not refer to any particular blends, to any particular properties or to any
particular applications.
[27]
It is worth noting that the expression “synergistically enhanced” is nowhere defined in the patent.
This is no doubt an important clue as to the significance of those words, as
one would have expected them to be defined somehow, if they were to be taken as
an explicit promise of a specific result. Indeed, Nova faults the Judge for not
having construed those terms, and for not having adopted the definition
provided by Dr. Speed in his expert report and in his oral testimony.
[28]
The common understanding of the “rule of mixing” is that the properties of a blend are
comparable to, and predictable from, the weight average ratios of the component
polymers. According to Dr. Speed, a “synergistic
response” means that a blend property is higher than, and a “destructive synergism” is lower than, the rule of
mixing. To borrow his own words:
A skilled person would understand
“synergistically enhanced” to mean that a property of a blend is better than
predicted by the rule of mixing and “destructive synergy” to mean the property
is worse than predicted by the rule.
Expert Report of Dr. Speed, Appeal Book Vol.
43, Tab 300 at para. 100
[29]
There are a number of problems with this
interpretation, not the least of which being that it is inconsistent with the
wording of the patent itself. In the parenthesis following the first use of the
words “destructive synergism”, the meaning
ascribed to that expression is rather that of a blend value that is lower than
either of the two pure components used to make the blend. It is true, as
pointed out by counsel for Nova, that Dr. Speed was cognizant of that definition;
but he dismisses it in a footnote on the basis that a skilled person would not
adopt this “literal meaning” (see Expert Report
of Dr. Speed, Appeal Book Vol. 43, Tab 300, p. 9231, para. 115, footnote 2).
[30]
Whatever the merits of the construction put
forward by Dr. Speed, the Judge was entitled to prefer the testimony of Dr.
Scott on that issue. The Judge quoted at length from Dr. Speed’s and Dr.
Scott’s expert reports, and came to the conclusion that Dr. Scott considered
the patent as a whole instead of focusing on two passages from the first page
of the patent. In particular, Dr. Scott looked at the examples outlined at
pages 16 to 26 of the patent, which show that the compositions defined by the
claims have improved properties, as compared to the primary type of polymers
previously used for film applications and the comparative heterogeneous polymer
blends. In summing up these comparisons, the patent speaks at page 26 of “good combination” of properties, “improvements”, and “higher
values”. According to Dr. Scott, this is not the hallmark of an explicit
promise of specific results, but rather a statement that the compositions of
the ‘705 Patent have improved on the prior art based polymers and polymer
compositions (see also the first paragraph of page 2 of the patent, quoted at
paragraph 8 of these reasons).
[31]
Dr. Scott also testified that the terms “synergy” and “synergistic”
are open to various definitions in the art and will take their colour from the
context into which they are used. Dr. Speed’s analysis of “synergistic enhancement” (i.e. a property of a blend
is better than predicted by the rule of mixing) is only one possible meaning,
and according to Dr. Scott would usually be referred to as “positive deviation from the rule of mixing”. As a
result of his underlying assumption that the patent is directed to his own
definition of “synergistic enhancement”, Dr.
Speed was led not only to disregard the parenthesis at the end of the second
paragraph at page 1 of the patent, but also to mischaracterize the data from
Table 3. The purpose of that Table was to compare blends of the invention with
the film properties of a heterogeneously branched interpolymer equivalent, with
comparable density and in some cases comparable melt index. Dr. Speed states
that incorrect comparisons have been made because the examples in the patent
are not directed at showing whether there is synergistic enhancement. However,
the ‘705 Patent was not directed to such synergistic enhancement. The
comparison is only meant to emphasize improvements in specific properties, not
in all properties and for all blends.
[32]
Such a reading is consistent with the trade-off
problem alluded to in the patent, according to which prior efforts to improve
one property of a polymer often resulted in another property being negatively
affected. The objective would therefore be to achieve a good balance of
properties, and to improve only those properties important to the particular
use contemplated. It should also be noted that nowhere does the patent compare
the properties of the inventive examples to the properties of the component
polymers of a blend, as would be required to demonstrate synergistic
enhancement as the term is defined by Dr. Speed; instead, the examples and the
Tables in the patent are meant to compare the inventive examples with polymers
having similar density and melt index (the two most important parameters in
evaluating the mechanical properties of polymers).
[33]
Dr. Scott’s construction of the patent was
certainly as credible and cogent as that of Dr. Speed, and the Judge could
certainly prefer his testimony over that of Dr. Speed. Moreover, Dr. Scott’s
construction of the patent is consistent with that put forward by Dr. Soares,
who is also of the view that there is no standard definition of the word “synergism”; in his Rebuttal Report, he stated that “[t]here is not a single ‘rule of mixing’ that applies to all
properties of polyolefin blends and the 705 Patent makes no promise regarding
such behaviour” (Rebuttal Report of Dr. Soares, Appeal Book Vol. 36, Tab
194, at para. 103). On cross-examination, he did not demur from that position (Cross-Examination
of Dr. Soares, Appeal Book Vol. 22, Tab 362, pp. 19755 to 19758).
[34]
As mentioned above, the Judge is entitled to
deference when assessing the evidence, and this is true for expert evidence as
much as for any other type of evidence: Bell Helicopter Textron Canada
Limitée v. Eurocopter, société par
actions simplifiée, 2013
FCA 219 at paras. 73-74, 449 N.R. 111; Corlac Inc. v. Weatherford Canada Inc.,
2011 FCA 228 at para. 24, 422 N.R. 49; Mylan Pharmaceuticals ULC v. AstraZeneca
Canada Inc., 2012 FCA 109 at para. 20, 432 N.R. 292; Wenzel Downhole
Tools Ltd. v. National-Oilwell Canada Ltd., 2012 FCA 333 at para. 44,
[2014] 2 F.C.R. 459; Zero Spill Systems (Int’l) v. Heide, 2015 FCA 115 at
para. 43, [2015] F.C.J. No. 554. The Judge was clearly alive to the conflicting
views of the experts with respect to the promise (or lack thereof) made by the
inventors in the patent, and he provided reasons for preferring the
interpretation given by Dow’s experts over that of Nova’s experts. Such a
finding is unassailable.
[35]
In a last-ditch
effort to convince the Court that the ‘705 Patent must be read as identifying a
specific promise that was described as being that “at
least impact strength will be better than the rule of mixing when other
properties like tensile strength, yield and modulus are surprisingly improved
in the newly discovered compositions”, Nova submitted that the patent
should be treated as a selection patent. For such a patent, it is necessary for
the specification to define in clear terms the nature of the characteristic
which the patentee alleges to be possessed by the selection for which he claims
a monopoly (Apotex Inc. v. Sanofi-Synthelabo Canada Inc., 2008 SCC 61 at
para. 114, [2008] 3 S.C.R. 265).
[36]
The problem with this theory is that there is no
evidence that this is how a skilled person would construe the patent. More
importantly, the Judge found that the ‘705 Patent was neither taught nor anticipated
by any of the two patents referenced by Nova. Since Nova has not appealed this
finding, it has no basis to argue that the patent is a selection patent and
therefore no basis to contend that an elevated promise of utility is required
to support the patent.
[37]
On the basis of the foregoing, I am therefore of
the view that the Judge could find that the inventors did not make an explicit
promise of a specific result, and that the patent did meet the test of a “mere scintilla” of utility. Nova has not seriously
challenged that finding, and one of its own experts acknowledged that the
invention solved the “trade-off problem”
(Cross-Examination of Dr. Brown, Appeal Book Vol. 15, Tab 347 at pp.
17015-17016). This ground of appeal is therefore dismissed.
2)
Did the Judge err in finding that the invention
was not obvious?
[38]
Nova argues that the Judge erred in dismissing
the allegation of obviousness. According to Nova, the patent discloses no
technical advance because blending all types of ethylene polymers was known and
common and therefore a claim to a new composition could not be inventive. This
argument was not pursued strenuously at the hearing, and rightly so.
[39]
First of all, it bears repeating that factual
findings made in applying the test for obviousness should not be reversed
absent a palpable and overriding error. In the case at bar, Nova does not
question the Judge’s identification of the applicable legal principles, and
must therefore meet this heavy onus of satisfying the Court that the Judge made
obvious and crucial errors in applying the law to the facts. Nova failed to do
so.
[40]
I pause to note that Nova’s argument is somewhat
disingenuous since Nova’s own patent application for its SURPASS product, filed
in 2002, asserts that blending different types of polymers is not free from
errors; the Judge quoted from that patent at paragraph 248 of his reasons.
Moreover, Nova’s own expert acknowledged the technical advance of the Dow
invention, stating that Dow’s invention overcame the “trade-off”
problem recognized in the prior art (see Cross-Examination of Dr. Brown, Appeal
Book Vol. 15, Tab 347 at pp. 17009 to 17013). One could stop there and dismiss
Nova’s argument on that sole basis. But there is more.
[41]
The Judge aptly pointed out that Dr. Speed
approached the obviousness issue from the wrong perspective. It is well
established since the decision of this Court in Beloit Canada Ltd. v. Valmet
OY (1986), 8 C.P.R. (3d) 289 at p. 294, 64 N.R. 287 that the reference for
the test of obviousness is not the competent inventor, but rather the
technician skilled in the art “having no scintilla of
inventiveness or imagination”. On cross-examination, Dr. Speed made it
clear that he approached the question of obviousness from the perspective of a
skilled person who has both inductive reasoning and imagination, an error that
the Judge identified and which accordingly undermined his analysis.
[42]
The Judge also refused to consider another
patent (the “Garza Patent”), published in April
1994, for the purpose of determining the obviousness of the ‘705 Patent, or
more specifically, the obviousness of Claim 11 of the ‘705 Patent. As
previously mentioned, the ‘705 Patent has a filing date of April 19, 1994 but
claims priority from a US patent application filed on April 28, 1993. Counsel
for Nova submitted that the US patent does not disclose the full density range
stated in part (a) of Claim 11, so that its claim date is therefore the filing
date of the ‘705 Patent. If that be the case, the Garza Patent should have been
considered for obviousness purposes with respect to Claim 11.
[43]
The Judge refused to decide the priority date of
Claim 11, however, for lack of evidence. He found that Nova failed to establish
that Dow could not claim the benefit of its April 1993 priority date for Claim
11. This is a pure finding of fact, and Nova cannot succeed on appeal by merely
restating the argument made at trial and rejected by the Judge without showing
a palpable and overriding error.
[44]
As a result, Nova’s submission must be
dismissed.
3)
Did the Judge err in finding that the claims
were not broader than the invention made or disclosed?
[45]
Nova submits that the Judge erred in finding
that the claims of the patent are not broader than the invention made or than
the invention disclosed in the specification. These are two well established
limitations on the extent of the monopoly which an inventor may validly claim (see
Farbwerke Hoechst A.G. vormals Meister Lucius & Bruning v. Canada
(Commissioner of Patents), [1966] Ex. C.R. 91 at para. 20, 31 Fox Pat. C.
64 (Ex. Ct.), aff’d [1966] S.C.R. 604, 50 C.P.R. 220 (S.C.C.)). The first
limitation is a question of fact, whereas the second is a question of
construction.
[46]
In his report, Dr. Speed suggested that he could
not find in the inventor’s documents any work showing that they made, or even
contemplated, compositions with more than one homogeneously branched
interpolymer or more than one heterogeneously branched interpolymer. Yet there
was evidence before the Judge that the inventors did contemplate compositions
with more than one Component A or B polymer. In any event, the Judge also
accepted Dr. Lai’s testimony that the invention only required Components A and
B, and that people were free to add other components to make further
improvements (Cross-Examination of Dr. Lai, Appeal Book Vol. 12, Tab 341 at pp.
15992-15994). In his expert report, Dr. Scott also stated that persons skilled
in the art would not require testing with other components as they would
appreciate from the experiments conducted on blends of Component A and Component
B polymers that additional polymers of this type could be included in the
blends without detracting from the overall benefit that the blend provides (Expert
Report of Dr. Scott, Appeal Book Vol. 35, Tab 192 at p. 6054, para. 168). That
view is consistent with the testimony given by Dr. Soares, according to whom
the POSITA would understand that other polymers in addition to the specific
claimed Component A and Component B could be included in the blends, provided they
do not adversely affect the properties of the composition (see the Rebuttal
Expert Report of Dr. Soares, Appeal Book Vol. 36, Tab 194 at para. 44; Examination
in Chief of Dr. Soares, Appeal Book Vol. 22, Tab 361 at pp. 19581 to 19582).
The Judge was entitled to prefer that evidence to that of Dr. Speed.
[47]
It is also clear that Dr. Speed’s view with
respect to the claims being broader than the invention disclosed was premised
on his theory that the ‘705 Patent promised an enhanced level of utility or “synergistically enhanced properties” for all
compositions. On that basis, he concluded that all claims are missing at least
one feature that the patent says is required of the interpolymers that are
useful for the claimed compositions. Having rejected that construction of the patent
and of its promise, however, the Judge was inexorably led to the conclusion
that Dr. Speed’s determination of what is essential and missing from the claims
was flawed. I see no error in such a finding.
[48]
It is no doubt true a claim may be too broad if
it gives the patentee a wider protection than his discovery entitles him to
receive. As President Thorson stated in Radio Corp. of America v. Raytheon
Manufacturing Co. (1957), 27 C.P.R. 1 at p. 22, [1956-60] Ex. C.R. 98 (Ex.
Ct.):
It is, I think, consistent with principle to
say that when a specification discloses the invention of a process for the
manufacture of an article in which the use of a special feature of the
invention is essential to the success of the invented process the inventor is
not entitled to claim a process for the manufacture of the article in which the
special feature is not used. He is not entitled to claim a monopoly more
extensive than is necessary to protect that which he has invented…
[49]
In the case at bar, Nova contends that Dow’s
claims are overbroad because they are missing defined limits as to “usefulness” in the patent. One such limit is that “the homogeneously branched ethylene/α-olefin
interpolymers do not contain a polymer fraction with a degree of branching less
than or equal to 2 methyls/1000 carbons”. Another is that “the homogeneously branched ethylene/α-olefin
interpolymers do not contain a polymer fraction with a degree of branching
equal to or more than 30 methyls/1000 carbons”. Since Claims 41 and 43
do not contain the first limitation and no claims contain the second
limitation, it is said that all the claims at issue are invalid for being
broader than the invention disclosed.
[50]
The Judge rejected that argument, not only
because it is tied to Dr. Speed’s flawed argument on an alleged promise of
synergy, but also because no evidence was presented that the person skilled in
the art would have considered any of these missing elements to be essential.
The fact that a feature is discussed in the specifications does not necessarily
mean that such a feature is of the essence of the patent (Whirlpool Corp. v.
Camco Inc. (1997), 76 C.P.R. (3d) 150 at p. 166, [1997] F.C.J. No. 1086
(F.C.T.D.); Lovell Manufacturing Co. v. Beatty Brothers Ltd. (1962), 41
C.P.R. 18 at p. 66, 23 Fox Pat. C. 112 (Ex. Ct.)). Yet, Dr. Speed seems to
assume that every feature of Components A and B referenced in the disclosure is
essential.
[51]
Both Dr. Soares and Dr. Scott opined that a
person skilled in the art would infer from the description that every preferred
range discussed is essential to both the making of the invention and their
understanding of how to make the blends of the invention. The Judge preferred
their evidence and accepted that the properties of the blend of polymers must
be looked at as a whole and not in isolation. This construction is bolstered by
the fact that some embodiments disclosed in the patent do not include the
allegedly essential features missing from some or all of the claims, which tend
to confirm Dr. Soares’ and Dr. Scott’s opinion that a person skilled in the art
would not consider these features as being essential. As a result, I can see no
palpable and overriding error in the Judge’s findings.
4)
Did the Judge err in inferring that SHC is
determined using a load/elongation curve?
[52]
In its Amended Statement of Claim, Dow argued
that the polyethylene film-grade copolymers manufactured and sold by Nova under
the SURPASS name met the limitations of Component A of all of the asserted
claims to the extent that they had an SHC greater than or equal to 1.3, thereby
infringing the ‘705 Patent. Critical to that argument is the procedure to
calculate the SHC. Nova asserts that the Judge erred in accepting Dow’s
submission that the SHC is to be determined by using load/elongation curves,
since the patent teaches that it must be done by using a stress/strain curve.
[53]
The patent teaches that improvements in the desired
properties of Component A can be predicted on the basis of the slope of strain
hardening coefficient. It adds that for the interpolymers used in the
invention, the SHC “is greater than 1.3, preferably
greater than 1.5”, and that it will typically be “less than 10, more typically less than 4, and most typically
less than 2.5”.
[54]
There is no dispute between the experts that the
skilled person with the common general knowledge as of 1994 would understand
the terms “strain hardening coefficient” in the
context of the ‘705 Patent to be a quantity related to the rate of change in
load with respect to elongation in the strain-hardening region, and the melt
index, which is a constant for a given polymer. The exact nature of that
relationship is set out at page 10 (lines 10-13) of the patent as: “SHC=(slope of strain hardening) * (I2)0.25, where I2 = melt
index in grams/10 minutes”.
[55]
The patent also teaches how the slope of strain
hardening is measured and the protocol for doing so. In particular, it states:
The tensile properties of the test sample is
tested on an Instron Tensile Tester at a crosshead speed of 1 inch/minute (2.5
cm/minute). The slope of strain hardening is calculated from the resulting
tensile curve by drawing a line parallel to the strain hardening region of the
resulting stress/strain curve. The strain hardening region occurs after the
sample has pulled its initial load (i.e., stress) usually with little or no
elongation during the initial load) and after the sample has gone through a
slight drawing stage (usually with little or no increase in load, but with
increasing elongation (i.e., strain)). In the strain hardening region, the load
and the elongation of the sample both continue to increase. The load increases
in the strain hardening region at a much lower rate than during the initial
load region and the elongation also increase, again at a rate lower than that
experienced in the drawing region. Figure 1 shows the various stages of the
stress/strain curve used to calculate the slope of strain hardening. The slope
of the parallel line in the strain hardening region is then determined.
‘705 Patent, p. 9, line 29 to p. 10, line 9
[56]
As noted by the Judge, the main dispute between
the parties on this issue boils down to the type of tensile curve that should
be used. Nova claims that the type of tensile curve to be used is an
engineering stress/strain curve, while Dow asserts that the type of tensile
curve to be used is a plot of the default load/elongation data from the Instron
Tensile Tester. Obviously, these two types of curve would yield different
numerical values.
[57]
Nova submits that the Judge erred by not
construing “stress/strain curve” with its plain
meaning as established in the common general knowledge on November 10, 1994.
Having considered the experts’ evidence and arguments of counsel, the Judge
found in favour of Dow and adopted its proposed construction of the patent that
SHC is measured using a load/elongation curve.
[58]
On appeal, Nova put forward much of the same
arguments submitted at trial. For example, Nova submitted that the Judge erred
by not construing the patent as a whole and by disregarding Example 3.
According to one of Nova’s experts (Dr. Cakmak), who undertook tensile tests
using the same apparatus referred to in the patent on various samples of Dow’s
products, no load/elongation curve would yield an SHC of 1.5, but some
stress/strain curves could. In Nova’s view, Example 3 should have informed the
meaning of “stress/strain curve”, and the Judge
erred in ignoring this point.
[59]
First of all, the Judge did not ignore that
submission but explicitly referred to it in his reasons (see Reasons for
Judgment, para. 93). The fact that he chose ultimately to prefer the evidence
of Dr. Young cannot be taken to mean that he forgot or misconceived Nova’s
argument. Both in his expert and rebuttal reports, Dr. Young makes it clear
that a skilled person would not undergo the sort of complex and lengthy
analysis performed by Dr. Cakmak, because the patent is self-contained and
directs a POSITA to do a load/elongation curve of a specimen of a particular
geometry. There would be no need to confirm this with an example, and Dr. Young
further provides in his report a number of reasons why a POSITA would
understand that example to be wrong.
[60]
Indeed, Dr. Cakmak himself acknowledged in his
report that a skilled person would not attempt to re-work Example 3 in an
attempt to construe the patent. He arrived at his construction of stress/strain
curve on other bases, and he appears to have considered that example mostly for
other purposes that are immaterial to this appeal.
[61]
Of more relevance, as noted by the Judge, is the
wording of the patent itself. In the excerpt reproduced above at paragraph 55,
the terminology of “load” and “elongation” is used repeatedly. I agree with Nova
that counting is no substitute for construing words. However, I disagree with
Nova’s contention that relating load to stress and elongation to strain (when
the patent says “load (i.e. stress)” and “elongation (i.e. strain)”) would necessarily be a
mere pleonasm if a load/elongation curve is to be used, since a skilled person
knows that stress increases proportionately to load and that strain increases
proportionately to elongation. As explained by Dr. Young, when associating load
with stress and elongation with strain, the patent could simply be defining “load” and “elongation”
as the specific types of “stress” and “strain” to be used. Nova clearly did not demonstrate
that the Judge made a palpable and overriding error in choosing that
construction over that suggested by Nova’s expert.
[62]
Such a reading would be consistent with Dr.
Young’s assertion that “stress/strain curve” is
often used generically to refer both to load versus elongation, stress versus
strain (or engineering stress versus engineering strain), and true stress
versus true strain, the evidence of which the Judge accepted. On appeal, Nova
submitted that this was an error because it was not the accepted meaning of
those terms at the date of the patent’s publication. I accept that Dr. Young
did not provide as many examples in the literature as he could have of the
generic use of “stress/strain curve”, but he did
affirm categorically more than once that such examples abound.
[63]
Moreover, Dr. Cakmak also accepted (after
initially denying) that the term “stress/strain curve”
could refer to other types of curves than an engineering stress/engineering
strain curve, after having been taken to a number of examples on
cross-examination where the term was used for other types of curves. This
wavering was noted by the Judge in his reasons (Reasons for Judgment, para.
87). It is no answer to say, as Dr. Cakmak claimed, that such use of “stress/strain” to refer to a load/elongation curve is
a misuse of that expression.
[64]
The Judge did not make much of Dr. Lai’s
evidence in his reasons, but he did mention that just like Dr. Young, Dr. Lai
uses the term “stress/strain curve” in a generic
way to cover different types of tensile curves including load/elongation
curves. Nova submits that this was improper, because Dr. Lai was not an expert
witness and could not give opinion on the construction of the patent. But the
Judge was clearly aware of Dr. Lai’s status, and the fact that he relied on
both Dr. Young and Dr. Lai in the same sentence is not sufficient to blur that
distinction; he clearly refers to Dr. Lai’s evidence to show that the generic
use of “stress/strain” curve was known and
accepted by skilled persons at the relevant period of time.
[65]
A further reason given by the Judge to accept
that the stress/strain curve is to be measured using a load/elongation curve is
that the output from the Instron Tensile Testing machine (which the patent
teaches is to be used to do the tensile testing) is load/elongation data and
the default output is a load/elongation curve. Nova counters that such an
inference could not be drawn since the method to convert load/elongation to
stress/strain was common general knowledge. This may well be so, and the Judge
acknowledges that such conversion is possible. However, that does not detract
from Dr. Young’s opinion that a skilled person would understand the patent to
refer to a load/elongation curve when stating that the slope of strain
hardening is calculated from the “resulting tensile
curve” (see ‘705 Patent at p. 9, lines 31-33), in view of the fact that
the output of the Instron machine is load and elongation data and a
load/elongation curve. The Judge clearly did not make a palpable and overriding
error in adopting that interpretation.
[66]
The last reason given by the Judge to accept
that the slope of strain hardening is to be measured using a load/elongation
curve in the context of the patent is the fact that no gauge length or
cross-sectional area is provided for converting load/elongation data to
engineering stress/strain values. That finding is supported by Dr. Young’s
evidence, according to which a standard gauge length would need to be specified
to create an engineering stress/strain curve. In the absence of such
specification, a skilled person would have many options in selecting a gauge
length to use for a specimen of the patent, and the gauge length chosen would
have a significant impact on the calculated engineering stress/strain values.
Nova’s own expert, Dr. Cakmak, did indeed consider two possible gauge lengths
for his experiments with Example 3 of the patent.
[67]
On appeal, Nova contends that a gauge length
would also be required for load/elongation measurements, and that Dr. Young
inferred what gauge length a skilled person would use from the specimen
geometry. I agree with counsel for Dow that this is a mischaracterization of
the evidence given by Dr. Young. Not only did he assert on a number of
occasions that gauge length is not required for load/elongation measurements,
but the value that he inferred from sample specifications described in the patent
were the grip separation and not the gauge length. While not explicitly making
that distinction in his reasons, the Judge defines “gauge
length” (at para. 89) in the portion of his reasons dealing with Nova’s
argument in this respect, and refers to “grip
separation” elsewhere in his reasons (see paras. 140 to 146, in
particular). I am therefore satisfied that the Judge appreciated the difference
between the two notions and did not make a palpable and overriding error in
coming to the conclusion, based on Dr. Young’s evidence, that the absence of a
gauge length value in the patent is further proof that the slope of strain
hardening is meant to be calculated using a load/elongation curve.
[68]
Nova’s last argument with respect to the
construction of the stress/strain curve is that the Judge ignored the evidence
of Drs. Cakmak and Speed that the common practice was to convert
load/elongation data to stress/strain to neutralize effects of dimensional
differences, and did not pay attention to Dr. Young’s use of a polynominal
correction formula to account for those differences. Indeed, Dr. Young
acknowledged that a skilled person would have known that the slope of strain
hardening should increase with specimen thickness (in other words, a thicker
specimen will require more force to elongate the same distance as a thinner
one). Consistently with his overall construction of the patent, however, he
chose to account for the differences in the specimen thickness by using a
mathematical formula instead of converting load/elongation data to
stress/strain.
[69]
It bears emphasizing that Dr. Young did not use
his correction formula to construe the patent, but only to determine whether
the tensile testing performed on Nova samples of its product substantiates an
infringement of the patent. In fact, and contrary to Nova’s submission, the
Judge deals with that correction formula in his reasons (at para. 139) as part
of his discussion relating to infringement. It also appears from the evidence
that Dr. Cakmak himself did in the past use a similar formula to fit data, and
that in appropriate circumstances it is a proper and reasonable approach.
Accordingly, I fail to see how the Judge made a palpable and overriding error
in accepting Dr. Young’s approach.
[70]
For all of the foregoing reasons, I find that
the Judge did not err in construing the patent and in inferring that SHC is
determined by using a load/elongation curve. Moreover, Nova does not contest
infringement of the SHC limitations if it is construed to refer to the load/elongation
curve. As a result, the Judge’s finding that the claims at issue in the patent
have been infringed by Nova’s Component A polymers of its SURPASS product, as
they all have an SHC value equal to 1.3 or greater, must be upheld.
5)
Did the Judge err in construing “heterogeneously branched”?
[71]
In his reasons, the Judge adopted Dow’s proposed
construction of the terms “heterogeneously branched”,
which is an essential descriptor of Component B in all claims, and defined that
expression in the negative (i.e. not homogeneously branched). He quoted the patent’s
definition of “homogeneously branched” and then
held “heterogeneously branched” to mean “a polymer having a distribution of branching different from
and broader than the homogeneously branched [polymer] at similar molecular
weight and SCB [short chain branching] averages”.
[72]
On appeal, Nova contends that the Judge erred by
assimilating the definition of “homogeneously branched”
found in the patent to the definition of “homogeneous
polymer” given by Dr. Soares, without appreciating that these are
different terms. As a result, the Judge erroneously found that Nova’s products
have a “heterogeneously branched” Component B
and therefore infringed all of the claims of the patent.
[73]
The parties agree that a polyethylene copolymer
is either homogeneously branched or heterogeneously branched. Since the
definition of “heterogeneously branched”
provided in the patent is dependent on the notion of “homogeneously
branched” (to the extent that it refers to a polymer having a
distribution of branching “different from and broader”
than the homogeneously branched polymer), the definition of “homogeneously branched” will obviously have a direct
impact on what is required of a polymer to be characterized as “heterogeneously branched”.
[74]
According to Nova, the Judge misconstrued the patent’s
definition of “homogeneously branched” polymer
and adopted too narrow a definition of those terms as a result of equating the
term “homogeneously branched” polymer with Dr.
Soares’ definition of “homogeneous polymer”
without appreciating the differences between those terms. In his expert report,
Dr. Soares defines a “homogeneous polymer” as
one made in the uniform polymerization conditions of a single-site catalyst in
an ideal CSTR (continuous stirred-tank reactor) that is perfectly well-mixed (see
Expert Report of Dr. Soares, Appeal Book Vol. 25, Tab 28 at paras. 37-41,
55-59, 75, 82-92 and 111-115). He added that such a polymer will display the
narrowest theoretically possible statistical distributions of molecular weight
and branching and will have an independent relationship between average
branching and molecular weight. While not disagreeing with that definition,
counsel for Nova submits that a “homogeneous polymer”
may be homogeneous in a number of ways, and that the claims only refer to “homogeneously branched” polymers, not to the narrower
category of “homogeneous polymer”.
[75]
In my opinion, this argument must fail. First, I
note that this construction of the patent was not put forward at trial.
Instead, Nova proposed that a Composition Distribution Branch Index (CDBI)
measurement should be used as a bright line test for distinguishing the two
types of polymers. The Judge easily rejected that proposal on the basis that
Nova’s own experts did not seem to agree as to where the bright line should be
drawn, and Nova did not challenge that finding on appeal.
[76]
Moreover, a careful reading of the patent
indicates that the two expressions are used interchangeably and that the use of
one as opposed to the other is of no critical import. It is also of some
interest that the patent, at page 12, mentions that heterogeneously branched
polymers differ from homogeneously branched polymers “primarily”
in their branching distribution, which seems to imply that they also differ on
the basis of other characteristics. In any event, the Judge was entitled to
prefer Dr. Soares’ construction over that of Dr. Mirabella in this respect.
[77]
Finally, Nova submits that the patent’s teaching
is inconsistent with the narrow definition of the words “homogeneously branched” espoused by the Judge. First,
the patent refers to various ranges of molecular weight distribution (MWD) for
a homogeneously branched polymer and all the claims specify the broadest of
these ranges; these go beyond the narrowest possible MWD one would expect for a
polymer made in the ideal conditions described by Dr. Soares and commonly known
as a “Flory distribution”. Second, the patent
does not require a homogeneously branched polymer to have the narrowest branching
distribution. And third, the patent does not mention an independent
relationship between average branching and molecular weight, which
characterizes a polymer made in the uniform conditions of a single-site
catalyst in a well-mixed reactor.
[78]
I need not address those arguments since the
Judge did not understand the words “homogeneously
branched” to mean a polymer made with a single-site catalyst in a
perfectly well-mixed reactor with the narrowest molecular weight and branching
distributions. Far from treating polymers made in those conditions as the
benchmark or the only polymers that can be considered “homogeneously
branched”, the Judge adopted the definition found in the patent itself
and applied it to the claims. It is true that a quote from Dr. Soares reproduced
by the Judge at paragraph 62 of his reasons may leave the impression that he is
using such a benchmark for comparison purposes; but the Judge did not rely on
this passage as providing a definition of “homogeneously
branched”.
[79]
Contrary to Nova’s submission, however, the patent
does mention an independent relationship between average branching and
molecular weight. The patent defines “homogeneously
branched” interpolymers as those “wherein
substantially all of the interpolymer molecules have the same
ethylene/comonomer ratio within that interpolymer”. In other words, the
average branching frequency (which is a measure of the ethylene/comonomer
ratio) does not change across the molecular weight distribution for a
homogeneously branched polymer. The relationship between branching and
molecular weight is therefore one of the criteria for distinguishing
homogeneous and heterogeneous polymers, and the Judge’s finding that this was
well known before 1994 is supported by the evidence and even by Nova’s own
witnesses (see Expert Report of Dr. Soares, Appeal Book Vol. 25, Tab 28 at
paras. 124-126; Cross-Examination of Dr. Kelusky, Appeal Book Vol. 14, Tab 345 at
pp. 16557-8 and 16566-7; Cross-Examination of Dr. Mirabella, Appeal Book Vol.
17, Tab 351 at p. 17788).
[80]
On the basis of the foregoing, I am therefore of
the view that the Judge made no palpable and overriding error in construing the
terms “heterogeneously branched” in the patent.
It follows that the Judge did not use the wrong test for finding that Nova’s
products have a “heterogeneously branched”
Component B.
6)
Did the Judge err in construing “comprising”?
[81]
Nova submits that the Judge erred in construing
the word “comprising” (found in all the asserted
claims) to mean “including, but not limited to”.
According to Nova, such a construction would mean that the claims of the patent
would be infringed by a composition consisting of as much as 85% of unclaimed
polymer(s), since the claimed minimums of Components A and B are respectively
10% and 5%. This is apparently the case of Nova’s SURPASS products, 50% of
which allegedly consist of unclaimed polymer. The construction of “comprising” put forward by Nova is that the
compositions consist “essentially” of Components
A and B, with “perhaps” small amounts of additives
as referred to in the patent at lines 14 to 20 on page 8. Nova finds support
for that construction from the fact that the examples provided in the patent
comprise only Components A and B, and that the outside limits in the claims are
set to total 100% of the compositions.
[82]
The problem with this reading of the patent is
that it is inconsistent with the ordinary meaning of the word “comprising”. As noted by the Judge, there is a
presumption of claim consistency, according to which the same words must be given
the same meaning throughout the claims and within any claim of a patent (see,
for example, Burton Parsons v. Hewlett-Packard (Canada) (1972), 7 C.P.R.
(2d) 198 at pp. 225-226, 1972 CarswellNat 531 (F.C.T.D.), rev’d 10 C.P.R. (2d)
126, [1973] F.C. 405 (F.C.A.), aff’d [1976] 1 S.C.R. 555, 17 C.P.R. (2d) 97
(S.C.C.); Johnson & Johnson Inc. v. Boston Scientific Ltd., 2008 FC
552 at paras. 208 and 213, 71 C.P.R. (4th) 123). In the case at bar, Nova’s own
expert (Dr. Speed) not only acknowledges that the traditional meaning of the
word “comprising” is “including
but not limited thereto”, but he also uses this traditional meaning in
his construction of that word as it appears within Claim 41 (‘705 Patent, p. 36,
line 14), though not in the first paragraph of that Claim (‘705 Patent, p. 35,
line 17). The Judge was certainly allowed to rely on that unorthodox submission
in preferring Dow’s more traditional interpretation of the word “comprising”.
[83]
Of course, Nova is correct to state that a
presumption should not be applied inflexibly as a rule, and that words must
take the colour of the context wherein they are used. That being said, I do not
think the Judge erred in rejecting Nova’s submission that in the context of the
‘705 Patent, “comprising” means that the compositions
must consist essentially of Components A and B, and perhaps the small amounts
of additives referred to at lines 14 to 20 on page 8 of the patent. There was
credible evidence from Dow’s expert witnesses to the contrary.
[84]
In his Rebuttal Report, Dr. Soares stated that a
skilled person would infer from the use of the word “comprising”
that the claims are not limited to Components A and B, but could include
additional polymers as long as those additional polymers do not deprive the
compositions of their desired benefits, either because of their nature, amount
or type. According to Dr. Soares, a skilled person would “very easily” arrive at that conclusion on the basis
of his/her knowledge of polymer blending and the intended use of the polymer
blend disclosed in the ‘705 Patent (Rebuttal Report of Dr. Soares, Appeal Book Vol.
36, Tab 194 at para. 44). Dr. Scott was of the exact same view (Expert Report
of Dr. Scott, Appeal Book Vol. 35, Tab 192 at para. 115). Indeed, Dr. Speed
himself equivocated on that issue and accepted at one point that a composition
containing Components A and B plus another Component C would infringe the patent
(Cross-Examination of Dr., Speed, Appeal Book Vol. 19, Tab 354 at pp.
18359-18360).
[85]
As for the Judge’s finding that at the very
least, Claim 11 is infringed irrespective of the definition given to the word “comprising”, there was expert evidence supporting
that finding. It is not in dispute that Component B of Claim 11 depends on Claim
10 and must therefore be 5 to 90 percent (by weight of the total composition),
have a density between 0.91 g/cc and 0.965 g/cc (as opposed to 0.93 to 0.965
for the other asserted claims), and be heterogeneously branched. It is not in
dispute either that the material in Reactor 2 meets the first and second
characteristics of Claim 11. The only point of contention between the parties
is with respect to the third requirement, that Component B must be
heterogeneously branched.
[86]
On appeal, Nova contends that its SURPASS
products do not infringe the specifications of Component B in Claim 11, because
the polymers in Reactor 2 are not heterogeneously branched. The Judge accepted
Dr. Soares’ testimony that the material in Reactor 2 is poorly mixed and
results in a continuum of materials which, taken as a whole, are
heterogeneously branched. Relying on its own expert, Dr. Mirabella, Nova
asserts that Dr. Soares does not take into account that there are two distinct
components in Reactor 2.
[87]
Once again, I am of the view that Nova has not
established a palpable and overriding error in the Judge’s finding. This is yet
another instance where this Court is asked to retry the case and second-guess
the Judge’s assessment of the evidence that was before him. Of course, a
different judge could have taken a different view and come to a different
conclusion on the basis of the same evidence; but this is not the test on
appeal.
7)
Did the Judge err in deciding that the HD
fraction was at least 5% by weight of the composition?
[88]
Nova also argues that the Judge erred in finding
an infringement of part B of all the claims. Dow had submitted that Nova’s
SURPASS polymers infringe all of the asserted claims of the ‘705 Patent if one
accepts Nova’s submission that the HD component made in the poorly mixed region
of its Reactor 2 as a distinct polymer. After noting that the parties initially
had three areas of disagreement in that respect (whether the HD component by
itself is heterogeneously branched, whether it is about 5% or more by weight of
the total SURPASS composition, and whether the HD component has a density from
0.93 g/cc to 0.965 g/cc), the Judge indicated that the third one was no longer
disputed in Nova’s closing submission but then dealt exclusively with the first
one without addressing whether the HD component met the second requirement.
[89]
This was clearly an oversight on the part of the
Judge, but it does not amount to a palpable and overriding error. While the
Judge should have dealt with this “area of
disagreement”, it was clearly not the most important one and there is sufficient
evidence to support a finding in favour of Dow on this narrow point.
[90]
The experts agree that it is not possible to clearly
separate the HD component from the other components found in Reactor 2. A
mathematical correction is therefore needed to separate the HD component for
each of the grades of SURPASS to determine whether they infringe the patent.
The disagreement between Dr. Soares and Dr. Mirabella relates to the correction
process required to obtain a cleaner separation between the higher density peak
(and the corresponding weight percent) and the contamination of the lower
density fraction. Dr. Mirabella testified that Dr. Soares did not follow his
own instructions and did not drop his correction line at the same place on
graphs showing the chemical composition distributions and profiles of the
polymers made in Dow’s ELITE product and Nova’s SURPASS products (at the left
of the “the valley between the two peaks” for
SURPASS products and “as close as possible to the
valley” for Dow’s ELITE product). As a result, it is submitted that Dr.
Soares overestimated the weight percentage of the HD component in Nova’s
SURPASS polymers.
[91]
Dow counters that Nova’s internal testing and
productions indicated that the HD component was above 5% by weight for each of
the representative product grades at issue, as required by the claims at issue
in the patent (see Expert Report of Dr. Soares, Appeal Book Vol. 25, Tab 28 at
paras. 212-216). Nova also admitted on discovery that the majority of the
resins had an HD component making up about 5% by weight of the product (see Discovery
transcript of Nova, Appeal Book Vol. 29, Tab 108 at pp. 2374 and 2382).
Finally, I note that Dr. Soares requested samples of all fractions collected at
Nova’s facilities for the fractionation experiments to test their contents but
Nova refused to provide them. We are therefore left with Dr. Soares’
explanation as to why Dr. Mirabella’s results with respect to the weight
proportion of the HD component are lower than 5%, which is that the high density
fractions that he used for his experiments did not in fact include all of the
HD components.
[92]
In light of this unchallenged evidence from Dow,
I am of the view that the failure of the Judge to address this issue was of no
consequence and does not undermine his overall finding that even if we only
consider the higher density component of Reactor 2, it still satisfies part B
of Claims 11, 29, 30, 33, 35, 36, 41 and 42. And in any event, as previously
discussed, the Judge did not make a palpable and overriding error in finding
that the whole polymer made in Reactor 2 satisfies part B of Claim 11.
IV.
Conclusion
[93]
For all of the foregoing reasons, I would
dismiss the appeal with costs. All of the arguments raised by the appellant
amount to no more than mere disagreements with the Judge’s factual findings and
assessment of the expert evidence. The Judge did not err in law, and was
entitled to prefer the evidence of some of the experts over that of others;
indeed, this is precisely the task he was called upon to fulfil, and the appellant
has not established that he made any palpable and overriding error in doing so.
“Yves de Montigny”
“I agree
Wyman W. Webb
J.A.”
“I agree
Richard Boivin J.A.”
Email this Content