PAULINE NEWMAN, Circuit Judge.

Howard A. Fromson appeals the decision of the United States District Court for the District of Massachusetts, ruling that the process for continuously anodizing aluminum used by Anitec Printing Plates, Inc. did not infringe Fromson's United States Patent No. Re. 29,754 (the '754 patent). The court also ruled that the '754 patent was invalid on the ground of obviousness. We affirm the holding of non-infringement, and reverse the holding of invalidity.

The Patented Invention

This invention relates to an improved process for continuously anodizing aluminum, whereby certain processing problems of electrical arcing and resistance heating are relieved, resulting in improved products and procedures accompanied by more rapid and efficient commercial production.

Metallic aluminum spontaneously oxidizes in air to form a thin coating of aluminum oxide upon its surface. This oxide coating protects the aluminum metal from further oxidation; however, the air-formed oxide is relatively soft, readily scratched and removed, and not of controllable durability and other properties that may be desired for specific uses. By the electrochemical process of anodization an aluminum oxide coating of controlled properties can be formed, by passing an electric current through a suitable oxygen-containing electrolyte, using the aluminum metal as the positive electrode in the circuit. Since, the positive electrode is the anode, the process is called anodization. The electrolytically formed aluminum oxide can be made to varying degrees of thickness, porosity, and hardness, selected for specific end uses of the anodized product. These properties are varied by adjusting the electrolytic, physical, and chemical conditions of the process in accordance with an extensive body of technical knowledge, experience, and skill.

The Fromson '754 patent is directed to a continuous anodization process wherein aluminum sheet or wire or other shapes, called the aluminum web, is passed through a tank containing anodizing electrolyte. More rapid, continuous processing increases production and reduces costs, but the faster the speed at which the aluminum web is passed through the anodizing cells the higher the electric current must be in order to produce an oxide layer of the desired thickness. In turn, the use of high electric current densities may produce arcing when mechanical contacts are used to charge the moving aluminum web, or may produce burning when electrolytic contact cells are used.

High current densities also generate high heat at the interface where the aluminum enters the electrolyte. Due to the low resistance at the interface, surging of electrical current as well as burning of the aluminum surface and possible boiling of the electrolyte can occur as the hot metal enters the cells. Another problem in commercial anodization processes is that of resistance heating, wherein the aluminum heats up as it conducts the anodizing current. Due to any of these events, the product may be marred and unacceptable for use. This combination of long-standing problems had not been satisfactorily solved at the time of Fromson's '754 invention, despite much effort over many years.

In the prior art it was known to conduct anodization by passing the aluminum web through an initial cell called a contact cell, before it entered the anodization cell. By using a contact cell to electrolytically introduce the anodizing current into the web, surface damage done by mechanical electrical contacts, such as brushes and rollers, was eliminated. According to the '754 invention certain problems of prior contact cell processes are ameliorated by directing the current flow so that the aluminum web is passed through an anodizing cell before it enters the contact cell, using two or more separate power sources. In the first cell an initial anodized oxide coating is produced; this coating protects the aluminum from surging current as it enters the contact cell and then a second anodizing cell. Anodizing current in the first cell is introduced by way of the cathodic contact cell, while a separate source of direct current connects the contact cell and the second anodizing cell. Additional cells and current sources may also be present. Thus the '754 process uses two or more direct current power sources, the first flowing between the first cell and the contact cell, the second flowing in the opposite direction between the contact cell and the second cell. The process is illustrated in Figure 2 of the '754 patent:

The claims in suit are 2, 11, and 13-15. Claim 2 is the broadest claim:

2. In a process for continuously electrolytically anodizing aluminum, the improvement which comprises introducing anodizing direct current from two or more sources in a cathodic contact cell having therein anodes connected to said sources of direct current, the aluminum having an anodized oxide coating formed thereon before entering said cell through the action of the direct current introduced in the contact cell itself.

Claim 11 specifically recites that the anodizing current is divided so as to flow through separate cross-sections of the web:

11. In a process for continuously electrolytically anodizing a moving aluminum web using the cathodic contact cell technique, the improvement which comprises introducing anodizing direct current from at least two direct current sources into said moving aluminum web into at least two cross-sections thereof, portion of the anodizing being carried out before the aluminum web enters the contact cell.

Claims 13-15 describe the process in greater detail. Claim 13, which depends from claim 11, recites that the anodizing currents flowing in two cross-sections of the web flow in opposite directions. Claims 14 and 15 both state that a porous anodic oxide coating is formed on the aluminum before it enters the contact cell. Claim 14 is illustrative:

14. Process for continuously forming a porous anodic oxide coating on aluminum which comprises:

(i) passing aluminum through a cathodic contact cell, said aluminum having a portion of the thickness of said porous anodic oxide coating already formed thereon before entering said contact cell through the action of anodizing direct current from a first source of direct current introduced into said contact cell; and

(ii) further anodizing the already anodized aluminum to form the remaining portion of the thickness of the porous anodic oxide coating by introducing anodizing direct current from a second source of direct current into said already anodized aluminum in said contact cell.

I [11] INFRINGEMENT

The Anitec process, as illustrated in an Anitec document at trial, also uses three cells, like the illustrated Fromson embodiment, plus a rinse step:

The district court construed the '754 claims whereby the oxide-forming action in Anitec's first electrolytic cell is not included in the meaning of "anodizing" and does not produce an "anodized" coating, as those terms are used in the claims. This was the major issue in dispute.

The claims define the patent grant and set its boundaries; that is, they establish the patentee's legal rights. When technical or scientific terms in the claims require definition or explanation or understanding in the course of deciding whether the claims are infringed, it is the judicial duty to do so. Markman v. Westview Instruments, Inc., 517 U.S. 370, ___, 116 S.Ct. 1384, 1387, 134 L.Ed.2d 577, 38 USPQ2d 1461, 1463 (1996) ("We hold that the construction of a patent, including terms of art within its claim, is exclusively within the province of the court.") In performing this construction, certain of the technologic facts were undisputed and others were strongly disputed insofar as technologic conclusions based thereon were material to the issue of the meaning and scope of the claims.

It was not disputed that Anitec uses a continuous electrolytic process with direct current flowing from two sources in a cathodic contact cell, and that the Anitec process forms an oxide coating on the aluminum in the first (phosphoric) cell before the aluminum enters the contact cell. Anitec conceded that its electrical connections, its cathodic contact cell, and its second (sulfuric) anodizing cell, are all within the meaning and scope of claims 2 and 11. There was no dispute that in the Anitec process the direct current was introduced from at least two sources into at least two cross-sections of the moving aluminum web, as these claims require.

However, Anitec argued that the first electrolytic cell in its process, although labelled by Anitec "Phosphoric Anodizing," is an etching and cleaning step, and that the formation of a thin layer of barrier-form aluminum oxide is not "anodization" as contemplated by the '754 patent. Anitec argued that its process does not, in the phosphoric cell, produce an "anodized oxide coating" (claim 2) or carry out a "portion of the anodizing" (claim 11) before the aluminum web enters the contact cell. Thus the question was whether the "anodized" coating that the '754 claims require before the aluminum web enters the contact cell includes within its scope the oxide coating formed in Anitec's "phosphoric anodizing" cell. The district court viewed the question as one of claim construction under Markman v. Westview Instruments, Inc., 52 F.3d 967, 34 USPQ2d 1321 (Fed. Cir. 1995) (en banc), aff'd, 517 U.S. 370, 116 S.Ct. 1384, 134 L.Ed.2d 577, 38 USPQ2d 1461 (1996).

Claim Construction

The analysis of infringement is a two-step progression. In the first step the court construes the claims of the patent; that is, the court explains and defines any disputed technical terms and sets the scope and boundaries of the claims. In so doing the court looks principally to the patent specification and any explanatory information in the patent examination record. The court may gain understanding of the state of the art at the time the invention was made, from the prior art or from experts in the technologic field. Physical and experimental evidence may be helpful, as in this case.

The district court held that the first anodizing step of the '754 process, before the web enters the contact cell, requires not only a porous oxide coating that is electrolytically produced on an aluminum anode, but a thicker layer of porous oxide as distinguished from the thin, non-porous oxide layer that is formed in the Anitec process. The district court relied primarily on the '754 specification which describes Fromson's first anodization step as producing porous oxide, on a Papst reference discussing formation of barrier oxide, and on findings based on extrinsic evidence derived from expert testimony, demonstrative evidence, and scientific tests. Fromson disputes this claim construction, stating that Anitec admitted that anodization occurs in its first electrolytic cell, that the plain text of the '754 claims reads on the Anitec process, and that the weight of the evidence supports Fromson's interpretation of the claims and their scope.

The Specification

To determine the meaning of disputed technical terms in claims, the first resource is the patent specification of which they are a part. Slimfold Mfg. Co. v. Kinkead Industries, Inc., 810 F.2d 1113, 1116, 1 USPQ2d 1563, 1566 (Fed. Cir. 1987) ("Claims are not interpreted in a vacuum, but are part of and are read in light of the specification.") In general, technical terms are deemed to have the same meaning in the body of the specification as in the claims. Patent specifications are written for persons of skill in the technical field, not for laymen, and often place the invention in the technologic context in which it must be understood.

The '754 specification requires that an anodized oxide coating be formed on the aluminum web before it enters into the contact cell, and teaches that the purpose of this preliminary coating is to protect the web from burning and current surging in the contact cell. The specification shows this coating to be porous oxide. While Fromson states, correctly, that a limitation from the body of the specification is not read into the claims unless the limitation is required to sustain patentability, the claims must be construed in light of the specification.

Although the '754 specification does not state that the thickness of the oxide, as the aluminum emerges from the first cell, must be greater than the thickness of naturally occurring oxide, the district court so interpreted the specification, concluding with the help of extrinsic evidence that for the coating to protect the web a thicker coating was required. The court resolved disputed testimony and documentary evidence to find that a thin barrier oxide layer enhanced burning rather than inhibited it, and that porous oxide was necessary at that stage. Thus the district court read the specification as supporting Anitec's position that since the oxide formed in its phosphoric cell is barrier oxide of under 5 nanometers in thickness, no thicker than naturally occurring oxide, the anodized oxide coating of the '754 claims can not be construed to include the Anitec process.

The Prosecution History and Prior Art

Fromson argues that since the aluminum web serves as the anode in Anitec's first cell, and an oxide layer is formed electrolytically, Anitec's process is anodization, whatever the structure of the oxide layer that is formed. The district court did not agree, despite the admission that Anitec itself consistently referred, in contemporaneous documents, to the process in the phosphoric cell as anodization. In reaching its conclusion the court relied not only on the specification, but on the prosecution of the Fromson patent as originally granted and during its reissue.

The prosecution history when a patent is reissued is part of the framework in which the patent is construed, as is the prosecution history of the original patent. Representations made and explanations presented during the reissue procedure may be relevant to interpretation of the claims. Fromson requested reissue of the patent after he learned of a Papst reference, German Patent No. 718,975, showing a two-cell anodization process in which the aluminum web enters the anodizing cell before it enters the contact cell. Fromson had originally broadly claimed this "reverse anodization" step. On reissue Fromson modified or cancelled certain broad claims, retaining the claims that contained the limitation of two or more separate direct current power sources and adding claims specific to formation of porous oxide in the first cell.

Fromson argues that the Papst reference was misapplied by the district court, for whether porous oxide was formed was not an issue in the reissue prosecution. Fromson states that he always intended to cover, and did cover in claims 2 and 11, anodizing to form either porous or non-porous oxide in the first cell. Fromson states that it was known that porous anodic oxide is formed only in a dissolving electrolyte such as sulfuric acid, and that barrier oxide was known to be formed in a non-dissolving electrolyte such as phosphoric acid. Fromson states that the Papst reference did not add to this knowledge, and that Papst did not circumscribe the Fromson teaching in the specification so as to exclude from the claims coverage of the formation of barrier oxide. Fromson states that Papst was significant only because of its teaching of what Fromson calls "reverse anodizing," that is, the anodization before entry of the web into the contact cell.

We take note that Papst states that its process can produce barrier oxide, and that Fromson disputed whether a web bearing barrier oxide can be used in further procedures such as galvanic deposition of metal. We conclude that the reissue prosecution does not of itself limit all of Fromson's claims to porous oxide, but that Papst's (and Fromson's) recognition that barrier oxide can be formed, along with the other evidence, supports the district court's ruling.

Extrinsic Evidence

Fromson argues that when "anodized" is given its standard definition, citing Webster's Third International Dictionary, of "to subject [a metal] to action by making [it] the anode of a cell before coating with a protective or decorative film," the Anitec process fully meets this definition. Fromson states that this standard definition is all that the claims require, and is supported by the specification and prosecution history. Fromson states that since the term has a plain meaning, the district court erred in relying on extrinsic evidence to change that meaning, citing Markman, 52 F.3d at 981, 34 USPQ2d at 1331 ("Extrinsic evidence is to be used for the court's understanding of the patent, not for the purpose of varying or contradicting the terms of the claims.")

Although Markman presents a useful general rule, it is adaptable to the needs of the particular case. In this case the technical experts not only aided the court's understanding of the technology, but they also provided evidence material to the interpretation of the claims. On this evidence the court enhanced the broadly stated dictionary definition of the process of anodization, with the understanding of the finer points of the technology that was provided by witnesses during the trial. Based on expert testimony, the court concluded that since not all electrolytically formed oxide coatings will protect the web as it moves into the contact cell — a fundamental purpose of the Fromson process — in the Fromson process the "anodized" coating before entry into the contact cell requires such a protective oxide.

Extrinsic evidence may be particularly helpful to the court when a specific technical aspect that is potentially of dispositive weight was not discussed in the specification or explored during the patent prosecution. This case illustrates the use of extrinsic evidence in order to determine the meaning and scope of a technical term as the term is used in the claims.

Witnesses explained that the web emerging from the first anodizing cell, unless sufficiently coated with porous oxide, would not be protected from current surging and burning in the contact cell and the second anodizing cell. Although Anitec's expert Barkman testified that it was known as early as 1958 that barrier oxide can prevent burning if precisely controlled, the district court gave greater weight to Anitec's evidence that the barrier oxide formed in Anitec's phosphoric cell did not have this effect. There was also presented analytical data that showed a thinning of the oxide layer while in the phosphoric cell, instead of a buildup of a thicker oxide layer.

There was also expert testimony and evidence based on tests varying the voltage in the phosphoric cell, showing that anodization as it occurred in the second anodization cell did not occur at the voltage used in Anitec's first cell, normally 8 volts. Other tests showed that line speed changes in Anitec's phosphoric cell did not change the oxide thickness of 4-5 nanometers, whereas line speed changes in the sulfuric cell produced proportional changes in oxide thickness, as is characteristic of the anodization process. There was evidence that the oxide formed in Anitec's first cell did not solve the burning problem, but exacerbated it. There was evidence that certain standard tests for anodization were not met for Anitec's barrier oxide (the eraser test, the bubble test, and the copper sulfate test). An expert witness testified that "anodization has different meanings in different contexts," in explaining Anitec's use of "anodization" for its phosphoric cell as contrasted with Fromson's use of the term in the '754 process.

The district court's findings of scientific/technologic fact were material to the issue of construction of the term "anodizing." The district court concluded that the clause in claim 2 "the aluminum having an anodized oxide coating found thereon" must be construed as requiring a thicker layer of oxide than that formed in air, and that it must be porous oxide. Claim 11 was similarly construed. The court held that the barrier oxide formed in the phosphoric cell was not "an anodized surface" as the term was used in the '754 claims:

I find that in 1973 no reasonable practitioner of this process would have had the opinion that a non-porous non-adherent oxide coating, as thin as the native 5 nanometer coating found naturally in the environment, of post-phosphoric oxide, constituted an anodized surface.

The district court found that "[t]he difference between a 1,000 or 2,000 nanometers oxide coating and a 5 nanometer oxide coating is not a difference in degree. It is a difference in kind. Something entirely different is going on." Thus the district court construed all of the claims in suit as limited to a process that produces an anodized coating in the precontact cell that achieves the purpose of protection against entry burning and resistance burning. The district court concluded as follows:

I conclude . . . that the terms "anodization," "anodized coating," and "anodized oxide coating" under Markman mean an electrolytically formed, adherent, porous aluminum oxide coating sufficiently thick (meaning thicker than native oxide) to, (A) protect against the burning and arcing problems addressed by the patent and optionally, (B) to provide a suitable porous base for other electrolytic treatments in the second anodizing cell such as electrophoresis, for metal plating.

Fromson readily admits that these purposes prompted his invention, but argues that the invention is not limited to these purposes, but is based on the use of a cathodic contact cell wherein an anodizing step occurs before entry into the contact cell. Thus Fromson argues that it is irrelevant to his process whether the oxide is porous or barrier, that either one can be used in the first step of the process. However, on all of the evidence, we agree that the district court correctly construed these terms in the claims.

Claim Differentiation

Fromson argues that the doctrine of claim differentiation prohibits interpretation of the broader claims of a patent so as to import into such claims the limitations of narrower claims. D.M.I., Inc. v. Deere Co., 755 F.2d 1570, 1574, 225 USPQ 236, 239 (Fed. Cir. 1985). Thus Fromson argues that since claims 14 and 15 specify the formation of a porous oxide on the aluminum web before it enters the contact cell, the doctrine of claim differentiation precludes importing this "porous" limitation into claims 2, 11, and 13. Fromson argues that this well-established doctrine precludes construing claims 2, 11, and 13 to require that the anodized coating be restricted to porous oxide.

The district court limited all of the claims in suit, treating as a matter of claim construction the issue of whether the claims cover, and thus would be infringed by, barrier oxide formation. Since the court was correct in its holding that claims 2, 11, 13 are not entitled, as a matter of law, to a definition of anodizing that embraces barrier oxide formation in the pre-contact cell step, this doctrine can not, of itself, restore to Fromson's claims a scope that has been shown to be unwarranted. See Tandon Corp. v. United States Int'l Trade Comm'n, 831 F.2d 1017, 1024, 4 USPQ2d 1283, 1288 (Fed. Cir. 1987) ("Whether or not claims differ from each other, one can not interpret a claim to be broader than what is contained in the specification and claims as filed.")

Literal Infringement

The district court construed "anodized" to mean an oxide coating having specific properties, as "anodized" characterizes the aluminum before it enters into the contact cell. The court held that the coating produced in the Anitec phosphoric cell is not "anodized" as the court construed this term. For the reasons discussed in connection with claim construction, we affirm the ruling that the Anitec process does not literally infringe the claims in suit.

This case illustrates the conceptual and practical overlap of "claim construction" as a matter of law, and the factual application of the claims to the accused process. The district court construed the '754 claims in the context of the anodized oxide formation in the phosphoric cell of the Anitec process. The extrinsic evidence on which the district court relied in its construction of the claims was geared to the infringement issue raised. Thus the court was able to limit its study to the facts placed in issue by the accused process.

We affirm the construction of the claims as correct. We do not discern clear error in the court's application of the claims, as construed, to the accused process.

The Doctrine of Equivalents

Infringement may be found under the doctrine of equivalents, although the claims as construed do not literally read on the accused process, if the accused process contains the same or equivalent elements or steps as the patented process. Warner-Jenkinson Co. v. Hilton Davis Chem. Co., ___ U.S. ___, ___, 117 S.Ct. 1040, 1054, 137 L.Ed.2d 146, 41 USPQ2d 1865, 1875 (1997). Thus infringement under the doctrine of equivalents may lie although the process is not within the literal scope of the claims as correctly construed.

Applying the criteria of Graver Tank Mfg. Co. v. Linde Air Prods. Co., 339 U.S. 605, 70 S.Ct. 854, 94 L.Ed. 1097, 85 USPQ 328 (1950), as modified in Warner-Jenkinson, the accused process must perform substantially the same steps as the patented process, in substantially the same way, to obtain the same result. Equivalency is a question of fact, and findings thereon are reviewed for clear error. Clear error exists when "the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed." Anderson v. City of Bessemer City, 470 U.S. 564, 573, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985).

Fromson argues that on the district court's claim construction there is infringement under the doctrine of equivalents. Fromson points out that the barrier oxide produced in Anitec's phosphoric cell is not natural oxide but the product of anodization, and that because it is harder and more adherent and protective than natural oxide it serves the function of an anodized oxide. Thus Fromson argues that the Anitec barrier oxide is equivalent to the Fromson porous oxide, for both are products of the same kind of electrolytic process and serve the same function in substantially the same way, to achieve the same result.

The district court did not agree. The court considered all of the evidence concerning the patented and the accused oxide coatings, and determined that the Anitec oxide formed in the phosphoric cell did not have the same properties as the Fromson oxide formed in the sulfuric cell, and did not serve the same purpose and solve the same problem. The court found that the primary function of the phosphoric cell in Anitec's process was to clean and condition the aluminum, not to provide an anodized layer to protect the web from surges or burning at high current densities. The court found that the thin barrier oxide coating initially formed on the Anitec web did not provide such protection, unlike Fromson's thicker porous coating. Thus the court found that there was not equivalency between the initial Anitec oxide and the "anodized coating" required by the '754 claims, in that the functions were not substantially the same. This finding has not been shown to be clearly erroneous.

The decision that there is not infringement under the doctrine of equivalents is affirmed.

II [52] VALIDITY

The district court held the '754 claims invalid on the ground that the invention would have been obvious in light of the prior art. Obviousness is a question of law based on underlying facts, Graham v. John Deere Co., 383 U.S. 1, 17, 86 S.Ct. 684, 693-94, 15 L.Ed.2d 545, 148 USPQ 459, 467 (1966). Included are the factual questions of the scope and content of the prior art, the differences between the claimed invention and the prior art, the level of ordinary skill in the field of the invention, and any objective evidence of unobviousness. Graham, 383 U.S. at 17-18, 86 S.Ct. 684, 693-94, 15 L.Ed.2d 545, 148 USPQ at 467; Glaverbel Societe Anonyme v. Northlake Mktg. Supply, Inc., 45 F.3d 1550, 1555, 33 USPQ2d 1496, 1499 (Fed. Cir. 1995). The determination of whether the subject matter as a whole would have been obvious to a person of ordinary skill at the time the invention was made, is reviewed for correctness as a matter of law. The underlying factual determinations are reviewed on appeal for clear error. Due consideration is given to the requirement that invalidation based on obviousness must be established by clear and convincing evidence. Id.

The parties generally agree as to the prior art and its teachings, and as to the level of ordinary skill in this field in 1972. The district court concluded that the patent was invalid "looking at Chubb, Cook, Barkman, and the other inventions that preceded this one. . . ." Fromson points out, and we must agree, that although much had been written on anodization, in this highly studied field no reference or combination of references showed or suggested the '754 process as a whole.

The Chubb reference describes producing an insulating coating of aluminum oxide on wire to be used in electrical machinery. Chubb shows a contact cell and the counter-current movement of the electric current, and teaches that "the electric current may be supplied to the conductor to be treated either before or after it has received an insulating coating and also in either the direction in which the wire is moving or in the opposite direction. . . ." This is quite different from the Fromson process. Other distinctions are that Chubb involves a single power source and shows only non-dissolving electrolytes (to produce a non-porous coating); a boilerplate statement that other "suitable liquids" can be used is insufficiently specific to include dissolving electrolytes such as are used by Fromson.

In the Barkman reference the contact cell precedes the pre-anodizing cell in the processing of the web, a process shown to have disadvantages that were remedied in the Fromson process. Although Anitec argues that Fromson's change in the sequence of steps is simply an "obvious choice," the choice apparently escaped Barkman, along with the benefits reported by Fromson. The Cook reference divides the anodizing cell into two or more compartments in order to avoid burning, but does not suggest use of an anodizing procedure before entry into the contact cell. The other references before the district court also contain differences from Fromson's process. They show the extensive effort expended in this area of technology, but none shows the process of the '754 patent.

That an inventor has probed the strengths and weaknesses of the prior art and discovered an improvement that escaped those who came before is indicative of unobviousness, not obviousness. The district court did not correctly apply the law of obviousness, for there is no suggestion or teaching in the prior art to select from the various known procedures and combine specific steps, along with a new electrical structure, in the way that is described and claimed by Fromson. The judgment of invalidity is reversed.

Damages

Damages and other issues raised on appeal are mooted by our affirmance of the judgment of non-infringement.

Costs

No costs.

AFFIRMED IN PART, REVERSED IN PART.