OSRAM GmbH v. International Trade Commission

United States Court of Appeals for the Federal Circuit

2006-1282

OSRAM GMBH,
and OSRAM OPTO SEMICONDUCTORS GMBH,

Appellants,

v.

INTERNATIONAL TRADE COMMISSION,

Appellee.

Alan D. Smith, Fish & Richardson P.C., of Boston, Massachusetts, argued for
appellant. With him on the brief were Charles H. Sanders and Christopher D. Agnew.

Michelle Walters, Attorney, Office of the General Counsel, United States
International Trade Commission, of Washington, DC, argued for appellee. With her on
the brief were James M. Lyons, General Counsel, and Wayne W. Herrington, Acting
Assistant General Counsel.

Appealed from: United States International Trade Commission
United States Court of Appeals for the Federal Circuit

2006-1282

OSRAM GMBH,
and OSRAM OPTO SEMICONDUCTORS GMBH,

Appellants,
v.

INTERNATIONAL TRADE COMMISSION,

Appellee.

__________________________

DECIDED: October 31, 2007
__________________________

Before NEWMAN, RADER, and DYK, Circuit Judges.

Opinion for the court filed by Circuit Judge NEWMAN. Dissenting opinion filed by Circuit
Judge DYK.

NEWMAN, Circuit Judge.

OSRAM GmbH and OSRAM Opto Semiconductors GmbH (collectively OSRAM)

appeal certain portions of the Final Determination of the United States International Trade

Commission (ITC or Commission) in an investigation conducted under Section 337 of the

Tariff Act of 1930 as amended, 19 U.S.C. '1337. 1 The patents subject of this appeal are

for a wavelength-converting composition wherein luminous phosphor particles convert the

1 In the Matter of Certain Light-Emitting Diodes and Products Containing Same,
Inv. No. 337-TA-512 (Int'l Trade Comm'n, Jan. 11, 2006) (Final Determination); Oct. 31,
2005 (Remand Initial Determination); Aug. 12, 2005 (validity and claim construction); May
10, 2005 (Final Initial Determination).
emitted light of light-emitting diodes (LEDs) to light of a different wavelength in order to

produce the desired white light. At OSRAM's request the ITC initiated an investigation

charging Dominant Semiconductors Sdn. Bhd. (and two other respondents no longer in the

case) with violating Section 337 by importing and selling compositions that infringe one or

more claims of OSRAM's patents called the "Particle Size Patents."

The Commission construed the claims of the Particle Size Patents, applied that

construction both to Dominant's accused products and OSRAM's domestic industry

products, and concluded that some of Dominant's imported products and all of OSRAM's

domestic products were not within the scope of the claims as construed. On this basis, the

ITC held that Section 337 was not violated.

We conclude that the ITC erred in its claim construction, and that on the correct

claim construction the Particle Size Patent claims are infringed and the domestic industry

prong of Section 337 is satisfied. The Commission's decision is reversed.

BACKGROUND

The Particle Size Patents are U.S. Patents No. 6,066,861 (the '861 patent); No.

6,277,301 (the '301 patent); No. 6,613,247 (the '247 patent); No. 6,245,259 (the '259

patent); and No. 6,592,780 (the '780 patent). The patents are directed to compositions,

methods, and uses wherein luminous pigment powders contain phosphors that produce a

spectral shift in the light emitted by electroluminescent components such as LEDs. The

phosphors absorb wavelengths in the ultraviolet, blue, or green ranges, and convert some

of the radiation to a higher wavelength, particularly in the yellow spectral range, whereby

the ensuing combination of complementary wavelengths appears white to observers.

2006-1282 2
 Light-emitting diodes are described as lasting longer than and using less energy

than traditional light sources, and the patented subject matter is described as overcoming

several disadvantages of prior products. The aspect of the claims relevant to this suit is the

grain size 2 of the pigment powders, claimed as having a maximum size of 20 micrometers

and a mean grain diameter of no more than 5 micrometers (μm) 3 . The meaning and the

measurement of this limitation are determinative of infringement. Claim 1 of the '861 patent

is representative, with the term at issue shown in boldface:

1. A wavelength-converting casting composition, for converting a wavelength
of ultraviolet, blue or green light emitted by an electroluminescent
component, comprising:
a transparent epoxy casting resin;
an inorganic luminous substance pigment powder dispersed in said
transparent epoxy resin, said pigment powder comprising luminous
substance pigments from a phosphorus [sic: phosphor] group having the
general formula A3B5X12:M, where A is an element selected from the group
consisting of Y, Ca, Sr; B is an element selected from the group consisting of
Al, Ga, Si; X is an element selected from the group consisting of O and S; M
is an element selected from the group consisting of Ce and Tb;
said luminous substance pigments having grain sizes # 20 μm and a
mean grain diameter d50 # 5 μm.

Dominant conceded that its imported powders meet all of the claim limitations except for

the "mean grain diameter d50 # 5 μm." Whether that limitation is met depends on how the

grain diameter is measured.

2 The terms "particle size" and "grain size" are used interchangeably by the
parties and in this opinion.

3 "μm" stands for micron or micrometer, denoting a millionth of a meter.

2006-1282 3
 In the first Initial Determination, the ALJ observed that the claims use the word

"mean," but with the symbol d50 whose conventional meaning is "median." Mean and

median do not always produce the same result, for "mean" is the average diameter, while

"median" is the diameter at which 50% of the particles are smaller and 50% of the particles

are larger. On this ground the ALJ held all of the claims invalid for indefiniteness. The full

Commission did not accept this ruling, and held that the claims can reasonably be

construed by application of the general rule that words prevail over symbols and that the

patentee can be its own lexicographer. The Commission explained that d50 is "a variable

defined by the words 'mean grain diameter' directly preceding it," and that the word "mean"

is used throughout the specifications and claims, whereas "median" does not appear in the

patents. Thus the full Commission concluded that "mean grain diameter d50" means the

mathematical average diameter of the grains, and rejected the ALJ's holding of invalidity on

the ground of indefiniteness.

The full Commission also deemed it unclear whether the mean grain diameter is

measured as the average diameter based on the number of grains, or the average

diameter based on the volume of the grains. This aspect of the claim construction had

evolved during the trial, as it became apparent that its resolution could be dispositive of

infringement. The Commission observed that the patent specifications did not state how

the mean diameter is determined, and selected the volume-based method; this is a

principal focus of this appeal. On this construction, the Commission remanded to the ALJ

for application to the products at issue. Applying the volume-based method, the ALJ found

that Dominant's "Fine Series LED" phosphors are within the claim limitation of having a

mean diameter of # 5μm, but that Dominant's "Normal Series LED" phosphors have a

2006-1282 4
volume-based mean diameter higher than 5μm. Thus the ALJ found that the Fine Series

products infringe the patents, but the Normal Series do not.

The ALJ also determined that OSRAM's own products of the domestic industry are

outside this claim limitation when measured by the volume-based method. The ALJ

concluded that OSRAM did not meet the "technical prong" of the domestic industry

requirement of Section 337, 19 U.S.C. '1337(a)(2). On this ground the ALJ ruled that

Section 337 was not violated as to any of the imported products. The full Commission

affirmed, and this appeal followed.

DISCUSSION

Rulings of the International Trade Commission are reviewed on the standard of the

Administrative Procedure Act, 5 U.S.C. '706(2)(E). 19 U.S.C. '1337(c). Rulings of law by

the ITC are reviewed for correctness, and findings of fact are reviewed to ascertain whether

they were supported by substantial evidence on the record as a whole. See Jazz Photo

Corp. v. International Trade Comm'n, 264 F.3d 1094, 1099 (Fed. Cir. 2001). Substantial

evidence is "such relevant evidence as a reasonable mind might accept as adequate to

support a conclusion." Consolidated Edison Co. v. N.L.R.B., 305 U.S. 197, 217 (1938).

Claim construction is reviewed as a matter of law. See Markman v. Westview Instruments,

Inc., 52 F.3d 967, 976 (Fed. Cir. 1995) (en banc), aff'd, 517 U.S. 370 (1996); Cybor Corp. v.

FAS Techs., Inc., 138 F.3d 1448, 1451 (Fed. Cir. 1998) (en banc).

I

No appeal is taken from the Commission's determinations that the claims are not

invalid for indefiniteness, or that the "mean grain diameter d50 " is the arithmetic average

diameter. The issues on appeal flow from the ruling that the average or mean grain

2006-1282 5
diameter of the phosphor grains is based on the volume, not the number, of grains.

OSRAM states that the mean or average diameter of the grains as set forth in the patents

would be readily understood by persons of experience in this field as the number-based

average, and that the Commission erred in choosing the volume-based average.

The number-based average is calculated as the sum of the diameters of all the

grains, divided by the number of grains. The volume-based average is calculated by

multiplying the diameter of each grain by its volume, summing the products thereof, and

dividing that sum by the sum of the volumes of the grains. These methods can produce

divergent results; OSRAM gives the example that by the number method the mean

diameter of a 1 μm grain and a 10 μm grain is 5.5 μm, whereas the mean diameter

calculated by the volume method is 9.99 μm 4 .

The Commission states that its choice of the volume method was "art-specific", and

thereby distinguished from the "general understanding" of how to describe the average

diameter of particles. The Commission cited two technical treatises: the Phosphor

Handbook and Perry's Chemical Engineers Handbook. The Phosphor Handbook discusses

the number-based method as generally used for phosphors, stating that the number-based

method "is easy to use, but both 'area-based' (volume-based) and weight-based methods

are frequently adopted to express the characteristics of actual powders." Perry's Chemical

Engineers Handbook states, in a general section for particle size designation, that "[i]t is

common to use a weight basis for percentage of frequency but surface or number may, in

some cases, be more relevant." The Commission placed primary reliance on these

4 Calculated by volume: ((1x13)+(10x103))/(13+103) = 9.99.

2006-1282 6
sources, plus the evidence that it is common for manufacturers of phosphors to report

phosphor size for sale to customers by volume. The Commission held that a person of

ordinary skill in this field would understand "mean grain diameter" to be based on the

volume method of measurement.

OSRAM states that this construction was incorrect, and contrary to the great weight

of evidence. The experts for both sides were in full and emphatic agreement that the

ordinary meaning of the average diameter of these particles is the number-based average -

- until Dominant's expert changed his position. OSRAM'S three expert witnesses testified

that average diameter of phosphors is generally measured by the number-based method,

and that the patents would be so understood by persons of ordinary skill in this field. This

testimony was not disputed. Dominant's expert testified unequivocally in his deposition that

"[t]o one skilled in the art, the term 'mean grain diameter d50' means that one determines

the mean or average grain diameter, which is defined as 'the integral (or sum) of the

diameter times the grain (particle) size distribution divided by the total number of grains

(particles);'" the witness later changed his position when it became apparent that this

question could determine liability. In addition, Dominant in its motion for summary

judgment had stated:

Here, the "mean grain diameter d50" refers to the average grain or particle
diameter, where "mean" is defined by its standard, well-known and accepted
meaning -- namely, the sum of the diameters of the grains or particles times
the grain size distribution divided by the total number of grains.

OSRAM's witness distinguished the way powders are sold from the way they are

characterized by scientists working on LED development, stating that "the R&D guy

[measures phosphor size] on a number basis only."

2006-1282 7
 The descriptive text in the patents is in accordance with the number-based

measurement, as both Dominant and OSRAM had presented it. The patent specification is

the primary resource for determining how an invention would be understood by persons

experienced in the field. See Phillips v. AWH Corp., 415 F.3d 1303, 1312-13 (Fed. Cir.

2005); Multiform Desiccants, Inc. v. Medzam, Ltd., 133 F.3d 1473, 1477 (Fed. Cir. 1998).

The patent specifications are in accordance with a number-based mean, for the invention

stresses the homogeneity of size, not volume, of the phosphor particles. It is not disputed

that homogeneity of size is better reflected in a number average than a volume average, for

a few large particles have a greater effect on the volume-based calculation. Expert witness

Dr. Zachau explained that "if you really said you want to do it on a volume basis, large

particles would be heavily overemphasized." Dr. Zachau explained that such distortion

arises because volume is calculated as the cube of the diameter, 5 and the cube of a larger

particle, for example a particle having a diameter of 10 μm, is a thousand times the cube of

a particle with a diameter of 1 μm. He explained that this cubing effect of the volume-

method obscures the information that particle size measurement is intended to convey, that

is, the average size of the particles.

As applied to the products at issue, OSRAM presents the example that for a powder

hypothetically containing 6,250 1-μm particles, 50 5-μm particles, and 1 20-μm particle, the

mean grain diameter calculated by the number-based method is 1.03 μm. In contrast, the

mean diameter calculated by the volume-based method is 9.6 μm. Dr Zachau testified that

5 Volume = 4/3 Jr3, where r = diameter/2

2006-1282 8
to achieve optimum homogenous light output, the number-based method provides the more

useful information:

For this context, [persons in this field] do it on a number basis only. . . . The
number distribution best measuresBis more appropriate for this application,
and for this device application. . . . Large particles, if they are very large, the
core doesn't even contribute because the light doesn't get fully in. . . . [T]he
small particles of the phosphor are very, very important, as we have seen
before, for the scattering . . . . So the small particles do have the most
important function here in the device. It's those we want to count. It's not the
large ones.

This testimony was not disputed by the witnesses for either side, although Dominant's

witness later amended his statement after it became clear that Dominant's position was

affected; the ALJ questioned the witness on this point, and it is noted that the ALJ did not

adopt the volume-based method in the Initial Determination.

We agree that the number-average measurement is better supported by the

specification, and that it provides the better description of a product whose purpose is

homogeneous distribution. See '861 patent, col.3, lines 21-25 ("[t]he luminous substance

pigments, with the above-indicated particle size, can advantageously . . . be dispersed

homogeneously in the epoxy casting resin."). When there is more than one method of

measurement and the patent does not explicitly discuss the methods, persons experienced

in the field are reasonably deemed to select the method that better measures the

parameters relevant to the invention. See Howmedica Osteonics Corp. v. Tranquil

Prospects, Ltd., 401 F.3d 1367, 1372 (Fed. Cir. 2005) (when a person of ordinary skill

would recognize which measurement is appropriate to the invention, that is the

measurement that applies). All of the experts agreed that the volume-based measure is

more sensitive to large particles, which do not function in the invention, and that the

2006-1282 9
number-based measure is more sensitive to the size and distribution of the particles that

perform the inventive function. There was no contrary evidence. Two commercial product

specification sheets that describe phosphor particles by mean volume and mean weight

had been submitted by OSRAM in the Patent and Trademark Office with its disclosure

documents, but not referred to by either OSRAM or the examiner during prosecution. It is

rare that references that were submitted with a disclosure document, but not even cited by

the examiner, are probative of an intent to depart from the plain technical meaning of terms

used in the specification and claims. See Boehringer Ingelheim Vetmedica, Inc. v.

Schering-Plough Corp., 320 F.3d 1339, 1347 (Fed. Cir. 2003) (references considered for

other purposes during prosecution do not establish that the patentee renounced the

ordinary meaning of a term as used in the specification and claims).

OSRAM states that the volume-based statement of particle size gives a less

accurate measure of the function of the LED, whereas the average diameter by number

better informs a person of ordinary skill whether the LED will operate well. Although the

Commission argues that the volume-based method, since more sensitive to "boulders," can

indicate whether boulders are present, OSRAM correctly states that the purpose of the

claim limitation is to state the parameters of the products that work in the desired way, not

those that may not. See Howmedica, 401 F.3d at 1372. The Commission erred in

construing the claims as requiring the volume-based method, contrary to the ordinary

meaning of the term as reflected in the specification and the testimony, and at odds with

the purposes of the invention. See Phillips, 415 F.3d at 1314 (claim terms are given the

meaning with which they are used in the patent specification); the court may consider

extrinsic evidence such as the testimony of experts in the field of the invention.

2006-1282 10
 This conclusion is reinforced by the undisputed fact that the volume-based measure

would exclude the OSRAM products that the patents were designed to cover. Cf. Hoechst

Celanese Corp. v. BP Chemicals, 78 F.3d 1575, 1581 (Fed. Cir. 1996) (a claim construction

that excludes the preferred embodiment is rarely, if ever, correct); Modine Mfg. Co. v.

United States International Trade Comm'n, 75 F.3d 1545, 1550 (Fed. Cir. 1996) ("a claim

interpretation that would exclude the inventor's device is rarely the correct interpretation").

We conclude that the Commission erred in construing "mean grain diameter" as the

volumetric mean, not the number-based mean. This ruling is reversed.

II

The ITC's ruling that Dominant's Normal Series LED products are not within the

claimed particle size was based on its application of the volume-based method. The record

before the ITC was that the average grain diameter of the Normal Series is significantly

above 5 μm 6 when measured by the volume-based method, and well below 5 μm when

measured by the number-based method. On the correct claim construction, Dominant's

Normal Series phosphor products are well within the literal scope of the claims.

The ITC states that if we should conclude that the correct measurement is indeed

the number-based method, infringement as to the Normal Series is unresolved because

that series is a "mixture" of two powders, one of which is outside of the claims before

mixture, although the mixture itself is within the 5 μm limitation. The Commission's position

is that "all of the pigments in the powder must together have a mean grain diameter d50

less than or equal to 5 μm." That is not the structure of the claims, which require that all

6 Dominant's actual particle sizes have been requested to be kept confidential.

2006-1282 11
particles be below 20 μm with the average diameter below 5 μm, but the claims do not

require that each of the separate powders that are mixed together meet the 5 μm limit

before mixing. The Commission did not respond to the record evidence showing, without

dispute, that the Dominant products include "mixed phosphor particles" having a number-

based mean grain diameter within the literal scope of the claims. On the Commission's

sketchy statement, and its incorrect view of the law, the requested remand for further trial

proceedings is without support.

On the correct construction that "mean grain diameter d50" is measured by the

number-based method, the ruling of noninfringement as to the Normal Series is reversed.

III

19 U.S.C. '1337 requires that the domestic industry meets certain criteria, in order

to invoke the right to exclude importation of infringing products. Section 337(a)(3) provides

that the requirement may be met when the patented inventions are the subject of, inter alia,

significant investment in plant and equipment by United States industry. See Texas

Instruments, inc. v. United States International Trade Comm'n, 988 F.2d 1165, 1180 (Fed.

Cir. 1993). The ITC concluded that OSRAM's investments in OSRAM Opto Inc. and

OSRAM Sylvania Inc. satisfy this requirement. However, the ITC found that OSRAM had

not shown that the phosphor particles so produced were covered by the Particle Size

Patents when measured by the volume-based method, that is, the method that the

Commission had designated as the correct measure of particle size.

The domestic product, to meet the technical prong test, Section 337(a)(2), must be

covered by the asserted claims; the test "is essentially the same as that for infringement,

i.e., a comparison of domestic products to the asserted claims." Alloc, Inc. v. International

2006-1282 12
Trade Comm'n, 342 F.3d 1361, 1375 (Fed. Cir. 2003). The Commission's ruling

concerning the domestic product was founded on the erroneous adoption of the volume-

based method, for the Commission observed that OSRAM showed only the number-based

mean particle size of its phosphors, whereas the Commission had ruled that the volume

mean is the correct measure. Thus the Commission held that OSRAM had not met its

burden of showing that it complied with the technical prong of Section 337. However, when

the domestic product's grain size is measured by the number-based method, it was

undisputed that the powders have a mean diameter below 5 μm. Dominant stated in its

post-trial brief that "Dominant does not challenge that OSRAM's products sold in the United

States are covered by the Particle Size Patents." OSRAM's evidence to this effect was

unopposed at trial.

On the corrected claim construction based on the number-based measure of mean

diameter, OSRAM's evidence was clear and unrebutted that the domestic product was

within the literal scope of the claims. The ITC's ruling that OSRAM did not meet the

technical prong of the domestic industry requirement is not supported by substantial

evidence on the record as a whole, and is reversed.

CONCLUSION

The ruling of noninfringement with respect to Dominant's Normal Series LED

products is reversed, as is the ruling that OSRAM did not meet the domestic industry

requirements. Thus, violation of Section 337 is established. We remand to the ITC for

proceedings consistent with this decision.

REVERSED and REMANDED

2006-1282 13
 United States Court of Appeals for the Federal Circuit

2006-1282

OSRAM GMBH,
and OSRAM OPTO SEMICONDUCTORS GMBH,

Appellants,

v.

INTERNATIONAL TRADE COMMISSION,

Appellee.

DYK, Circuit Judge, dissenting.

I respectfully disagree with the majority’s construction of the term “mean grain

diameter d50,” which is included as a limitation of the asserted claims in the five patents

at issue here—patents which the majority designates the “Particle Size Patents.” 1

In my view, the International Trade Commission (“ITC”) reached the correct claim

construction. Accordingly, I would affirm the ITC’s holding, based on its claim

construction, that the “Normal Series” products of the accused infringer, Dominant

Semiconductors Sdn. Bhd. (“Dominant”), do not infringe the Particle Size Patents. I

also would affirm the ITC’s holding, based on its claim construction, that appellants

(collectively “OSRAM”) do not practice the Particle Size Patents in their domestic

1
The disputed language is present in each of the claims of the Particle Size
Patents that appellants assert were infringed: claims 1, 3, 6-7, and 10-13 of Patent No.
6,066,861; claims 1-2, 6-7, 11-12, and 14-15 of Patent No. 6,277,301; claims 1, 3, 6-7,
10-15, 17, and 20-21 of Patent No. 6,613,247; claims 1, 3, 6-7, 10-13, and 15 of Patent
No. 6,245,259; and claims 2-5, 7, and 10 of Patent No. 6,592,780.
operations and therefore have failed to make the showing required by the domestic

industry prong of section 337 of the Tariff Act of 1930, as amended, 19 U.S.C. § 1337.

The Particle Size Patents relate to the use of a phosphor powder in a light-

emitting diode (“LED”) to transform some light emitted by the LED from one wavelength

to another to give the appearance that the LED emits white light. Relatively small

phosphor grains absorb light of one wavelength and emit and scatter light of a different,

complementary wavelength to produce the appearance of white light. Larger particles,

however, are detrimental to the invention. They do not function in the same way as

smaller particles to absorb and emit light and are not as effective at scattering light to

produce a uniform color and intensity. Larger particles also cause problems related to

uneven sedimentation during the manufacturing process.

As the majority recognizes, Maj. Op. at 5-6, there are two possible methods to

calculate the “mean grain diameter d50” of a pigment powder. The first is an average

diameter by number of particles, and the second is an average diameter by volume.

Neither the claims nor the specifications of any of the Particle Size Patents state which

of the two methods is to be used.

Contrary to the majority’s assertion, the expert witness testimony did not reflect

“full and emphatic agreement that the ordinary meaning of the average diameter of the[]

particles is the number-based average.” Maj. Op. at 7. Rather, the record indicates that

both methods were used in the industry for differing purposes. When phosphor

powders were sold, including by OSRAM, the particle size was measured as an

average diameter by volume. J.A. at 3404, 4318-19, 4330. When conducting research

and development of new products, researchers employed a number-based average to

2006-1282 2
measure particle size. J.A. at 4330. Despite the majority’s suggestion that Dominant’s

expert witness changed his testimony on this point, there is in fact no conflicting

testimony. Indeed, OSRAM’s employee and expert, Dr. Zachau, testified that the

volume-based average particle diameter is used commercially in the sale of phosphor

powders, while the numerical average diameter is used for research and development

purposes. Id. He explained that the measurement required to compute the average

diameter by volume is “an easy, fast, an [sic] inexpensive measurement,” but that a

numerical average diameter must be calculated from less efficient measurements that

require the use of a scanning electron microscope. Id. The central issue here is

whether the “mean grain diameter d50” should be defined from the perspective of

commercial sales or research and development.

In my view, the ITC was correct to choose the commercial sales definition—that

is, an average by volume—as the correct construction of the “mean grain diameter d50.”

The specifications here make it quite clear that a commercial invention is being

described. See, e.g., J.A. at 114 (describing, as an object of the invention, to “enable[]

mass production at reasonable engineering effort and expense and with maximally

replicable component characteristics”). There was objective evidence of the

widespread commercial use of this calculation method, and undisputed evidence that

the measurements required are relatively cheap and efficient, as would be required for

commercial use. Those who secure patents typically are describing devices and

methods designed for commercial use, rather than devices designed only for research.

It follows, I believe, that the patent should be interpreted to utilize this commercial

measurement rather than the research measurement.

2006-1282 3
 The ITC also properly relied on two technical treatises, the Phosphor Handbook

and Perry’s Chemical Engineers Handbook. See Phillips v. AWH Corp., 415 F.3d 1303,

1318 (Fed. Cir. 2005) (en banc) (noting usefulness of technical treatises in construing

claims). While those treatises do not define the term “mean grain diameter,” the ITC

concluded that both “indicate that the weight basis is more often used to describe real

powders than other bases.” J.A. at 48. Weight- and volume-based average particle

diameters are essentially equivalent because they are related based on a known

constant, the particles’ density. The commercial sales literature and the technical

treatises both constitute particularly strong sources of extrinsic evidence under the

circumstances of this case because they provide objective, contemporaneous,

unbiased, and publicly available descriptions of how mean particle size was measured

by those skilled in the art. See Phillips, 415 F.3d at 1322 (describing proper use of

contemporaneous extrinsic evidence, especially evidence from an “unbiased source

‘accessible to the public in advance of litigation’” (quoting Vitronics Corp. v.

Conceptronic, Inc., 90 F.3d 1576, 1585 (Fed. Cir. 1996))).

The majority reasons that the purposes of the invention described in the Particle

Size Patents are better served by a numerical average, because the majority assumes

that the disputed claim language sought to emphasize the prevalence of small particles

rather than the absence of large particles. 2 The difference between an average

2
The majority also relies on the fact that OSRAM’s own products do not fall
within the claim limitation when measured by a volume-based average particle size.
Maj. Op. at 11. In relying on this court’s opinion in Hoechst Celanese Corp. v. BP
Chemicals, 78 F.3d 1575, 1581 (Fed. Cir. 1996), however, the majority mistakenly
equates OSRAM’s products with the preferred embodiment of the patent. See Int’l
Visual Corp. v. Crown Metal Co., 991 F.2d 768, 771-72 (Fed. Cir. 1993) (reversing claim

2006-1282 4
diameter based on the number of particles and an average diameter based on volume

is that the former tends to emphasize the presence of many useful small particles, while

the latter gives greater emphasis to the presence (or absence) of any larger,

undesirable particles. The majority’s conclusion that the concern was with the number

of small particles is unsupported by any language in the claims or specifications of the

Particle Size Patents, and in my view is entirely speculative. In other words, there is no

intrinsic evidence to suggest that the patents sought to emphasize the presence of

many useful small particles, rather than the relative absence of larger, harmful particles.

Under these circumstances, it is just as likely that the disputed claim language sought to

emphasize the absence of larger particles. One of OSRAM’s experts explained that in

the process of invention, larger particles proved harmful to both the manufacturing

process and the function of the final product, and that OSRAM obtained better results

by using smaller particles. J.A. at 5041-43.

The majority’s contention that a claim limitation necessarily seeks to “state the

parameters of the products that work in the desired way,” rather than specifying in the

negative parameters that have proven detrimental, Maj. Op. at 10 (citing Howmedica

Osteonics Corp. v. Tranquil Prospects, Ltd., 401 F.3d 1367, 1372 (Fed. Cir. 2005)), is

construction based on commercial embodiment because “’[i]nfringement is determined
on the basis of the claims, not on the basis of a comparison with the patentee's
commercial embodiment of the claimed invention.’” (quoting ACS Hosp. Sys., Inc. v.
Montefiore Hosp., 732 F.2d 1572, 1578 (Fed. Cir. 1984))); see also SmithKline
Beecham Corp. v. Apotex Corp., 403 F.3d 1331, 1339 (Fed. Cir. 2005) (rejecting claim
interpretation based on commercial embodiment of invention). At oral argument,
OSRAM was unable to identify any record evidence establishing that its products were
designed to practice the Particle Size Patents or that the preferred embodiments under
the patents would not be within the claim limitation if the average particle size were
measured as a volume-based mean.

2006-1282 5
unsupported as a matter of law and contrary to common sense. This court’s opinion in

Howmedica did not discuss whether claim limitations properly are interpreted as

describing either virtuous qualities or the avoidance of undesirable qualities. In

Howmedica, there were two possible methods to measure the “transverse sectional

dimensions” of a stem part designed to secure a prosthetic limb to a bone socket. 401

F.3d at 1371. This court reasoned that the method relying on two-dimensional surface

area was the appropriate construction because in several places the specification

indicated the need for the stem part to fit closely into the bone socket, and the two-

dimensional surface area method would produce a more exact fit than the other

method. Id. at 1372. Thus, while the two-dimensional surface area measurement

method in Howmedica was directed to specifying a parameter with useful effects, there

was no indication that the rejected method sought to avoid a parameter with detrimental

effects.

Here a different situation prevails, and the purposes of the invention could be

served either by utilizing the volume-based or the number-based measurement. Under

these circumstances, the objectives of the invention do not answer the claim

construction question, and resort to the commercial standard of a volume-based

average is appropriate. I dissent from the majority’s refusal to sustain the ITC’s

adoption of the commercial standard.

2006-1282 6