Opinion for the court filed by Circuit Judge NEWMAN for Parts I and II; dissenting opinion as to Part I filed by Chief Judge ARCHER. For Part III, Chief Judge ARCHER and Circuit Judge MAYER each filed opinions concurring in the judgment of affirmance; dissenting opinion as to Part III filed by Circuit Judge NEWMAN.
PAULINE NEWMAN, Circuit Judge.Advanced Semiconductor Materials Amer-ica, Inc. and Epsilon Technology, Inc., together doing business as ASM Epitaxy (herein “ASM”), appeal those portions of the decision of the United States District Court for the Northern District of California1 sus-*1566taming the validity of United States Patent No. 4,496,609 (the ’609 or McNeilly patent), and holding that United States Patent No. 4,728,389 (the ’389 or Logar patent) is infringed by application of the doctrine of equivalents. Applied Materials, Inc., owner of the ’609 and ’389 patents, appeals the district court’s holding that its United States Patent No. 4,081,313 (the ’313 patent) is invalid and that the ’389 patent is not literally infringed. The validity of the ’389 patent was not contested. The ruling that the ’609 patent is infringed in not appealed.
We affirm the decisions that the ’609 patent is valid and that the ’389 patent is not literally infringed, reverse the decision that the ’389 patent is infringed under the doctrine of equivalents, and affirm the decision that the ’313 patent is invalid. The case is remanded for appropriate further proceedings.
DISCUSSION
Both Applied Materials and ASM are manufacturers and sellers of devices used in the production of semiconductor integrated circuits by chemical vapor deposition (CVD). By the process of CVD thin layers of chemical materials such as silicon and silicon compounds are deposited upon a substrate by exposing the substrate at high temperatures to the vaporized material to be deposited, under controlled conditions in oven-like reactor chambers. Due to the extremely small size and complex detail of the circuits produced and the thinness of the layers deposited, the control of contaminants during CVD is of paramount importance. The inventions of the patents in suit are directed to improvements in the methods and apparatus of CVD for the purposes of reducing or eliminating contaminants and enhancing the quality of the deposition product.
I
THE ’609 (McNEILLY) PATENT
The purity of the deposition is affected by contaminants emanating from or passing through the quartz walls of the deposition chamber, and by the migration of contaminants onto the substrate by gravity or static electricity. A known method of reducing the contamination that arises from the quartz walls is to keep the walls cool during the deposition. Since the CVD substrate is heated to very high temperatures during the deposition, generally over 1100° C, the heating of the substrate must be accomplished without heating the walls of the reaction chamber. The invention of the ’609 patent is directed to an improved process for heating the substrate in a quartz reaction chamber, using radiant heat energy.
The prior CVD process of choice used radio frequency energy as the source of deposition heat. Since the quartz chamber walls transmit but do not absorb radio frequencies, the walls remain cool and contamination is reduced. The substrate does not absorb radio frequency energy, and is heated indirectly by placing it on a slab of graphite called a susceptor; the susceptor absorbs the radio frequency, converts it to heat, and heats the substrate by contact.
The invention of the ’609 patent is the use of high intensity heat lamps and radiant energy to heat the substrate, instead of using radio frequency energy. The prior art recognized this general use of radiant energy, but stated that it was not effective at high deposition temperatures. McNeilly and coworkers discovered not only that radiant energy can be used at high temperatures, but that it resulted in unexpected advantages in that it produced more uniform heating of the substrate and superior circuitry. The district court observed that after initial skepticism the Applied Materials method, subject of the ’609 patent, was recognized as superior by the industry, and that Applied Materials’ reactor has displaced the radio frequency reactor as the CVD apparatus of choice.
Applied Materials charged ASM with infringement of the ’609 patent by ASM’s Epsilon One reactor. The district court rejected ASM’s defenses of patent invalidity and un-enforceability, and found that the ’609 patent was infringed. ASM appeals the validity determination on two grounds: double patenting and obviousness. Other issues concerning the ’609 patent are not appealed.
*1567 Double Patenting
ASM argues that the ’609 patent is invalid for double patenting with United States Patents No. 8,628,712 (the ’712 patent) and No. 4,047,496 (the ’496 patent). These three patents all arose from the same parent application. The ’712 and ’496 patents claim the apparatus and the ’609 patent claims the process for this use of radiant energy. Claim 1, the broadest claim of the ’609 patent, is shown in the margin.2
The question of whether the claims of the process patent are subject to double patenting with the claims of the apparatus patents arises only if these patents are not entitled to the benefit of 35 U.S.C. § 121. ASM argues that Applied Materials lost this entitlement because during prosecution before the Patent and Trademark Office (PTO) the applicant enlarged the process claims beyond their scope at the time of imposition of a “restriction requirement.” ASM states that this action removed all of the ’609 claims from the benefit of § 121, and requires that the ’609 patent be either invalidated for double patenting or subjected to a terminal disclaimer. These issues relating to § 121 are matters of statutory interpretation, to which we give plenary review on appeal.
The patent examiner imposed a restriction requirement on the McNeilly patent application as it was originally filed, in accordance with the first sentence of 35 U.S.C. § 121:
§ 121[1] If two or more independent and distinct inventions are claimed in one application, the Commissioner may require the application to be restricted to one of the inventions.
As required by administrative rule, in imposing the restriction requirement the examiner divided the claims into groups, viz. “an oven-type radiation heated reactor,” “a reactor with means for introducing gaseous reactants,” and “a gaseous epitaxial coating process.” In response to the examiner’s requirement that the applicant elect which invention he wished to continue prosecuting in the original application, McNeilly elected the radiation-heated reactor and duly prosecuted the other two inventions in separately filed divisional applications. The second sentence of § 121 recognizes this procedure and provides that such divisional applications are entitled to the benefit of the original filing date:
§ 121[2] If the other invention is made the subject of a divisional application which complies with the requirements of section 120 of this title it shall be entitled to the benefit of the filing date of the original application.
The three patent applications issued as the ’712 and ’496 patents on the apparatus and the ’609 patent on the process. We take note that the history of these patents shows several refilings, amendments, and continuations-in-part, as discussed in Applied Materials, Inc. v. Gemini Research Corp., 835 F.2d 279, 15 USPQ2d 1816 (Fed.Cir.1988).
The process claims as presented in the original application and carried into the ’609 divisional application were drawn to the deposition of epitaxial layers.3 During further prosecution the claims were amended to include chemical vapor deposition broadly, thus encompassing non-epitaxial deposition as well as epitaxial, as generally described in the specification. The district court observed that the PTO classification, a criterion of restriction requirements, was unchanged. ASM argues that this change in scope of the claims after the restriction requirement was *1568initially imposed eliminates the benefit of the third sentence of § 121:
§ 121[3] A patent issuing on an application with respect to which a requirement for restriction under this section has been made, or on an application filed as a result of such a requirement, shall not be used as a reference either in the Patent and Trademark Office or in the courts against a divisional application or against the original application or any patent issued on either of them, if the divisional application is filed before the issuance of the patent on the other application.
ASM states that without the shield of § 121 the ’609 process patent is invalid for double patenting with the ’712 and the ’496 apparatus patents. ASM states that this shield is lost if there is any change in the scope of the claims, whether or not the PTO classification is unchanged.
The purpose of the rule against double patenting is to prevent an inventor from effectively extending the term of exclusivity by the subsequent patenting of variations that are not patentably distinct from the first-patented invention. See In re Vogel, 57 C.C.P.A. 920, 422 F.2d 438, 164 USPQ 619 (1970) (discussing the judge-made law of double patenting). Thus a patentee can not obtain a later patent for the same invention that has already been patented. Nor can a patentee obtain a later patent for an obvious variant of the invention claimed in the earlier patent, unless the patentee disclaims the term of the later patent beyond the expiration of the earlier one. See Quad Environmental Technologies Corp. v. Union Sanitary District, 946 F.2d 870, 873, 20 USPQ2d 1392, 1394 (Fed.Cir.1991); In re Longi, 759 F.2d 887, 892, 225 USPQ 645, 648 (Fed.Cir. 1985).
However, when the existence of multiple patents is due to the administrative requirements imposed by the Patent and Trademark Office, 35 U.S.C. § 121 provides that the inventor shall not be prejudiced by having complied with those requirements. Thus when two or more patents result from a PTO restriction requirement, whereby aspects of the original application must be divided into separate applications, § 121 insulates the ensuing patents from the charge of double patenting. See Studiengesellschaft Kohle mbH v. Northern Petrochemical Co., 784 F.2d 351, 354, 228 USPQ 837, 840 (Fed.Cir.), cert. dismissed, 478 U.S. 1028, 106 S.Ct. 3343, 92 L.Ed.2d 763 (1986).
The district court held that the amendments to the process claims during subsequent prosecution of the divisional ’609 application did not eliminate the benefit of § 121 from the ensuing patent. A restriction requirement does not prohibit subsequent amendments to the claims. As discussed in Texas Instruments, Inc. v. U.S. Int’l Trade Comm’n, 988 F.2d 1165, 1179, 26 USPQ2d 1018, 1029 (Fed.Cir.1993), the examiner’s demarcation among the separate inventions should be preserved. However, even if such consonance is lost, double patenting does not follow if the requirements of § 121 are met or if the claims are in fact patentably distinct. See id. at 1179, 26 USPQ2d at 1029-30. In this case consonance was not violated, for the process claims remained in separate patents from the apparatus claims although the scope of the process claims was modified.
ASM argues that the restriction requirement might not have been imposed at all if the non-epitaxial process claims had been present initially. However, the benefit of § 121 is not lost simply because it might not have been imposed. The purpose of § 121 is to accommodate administrative convenience and to protect the patentee from technical flaws based on this unappealable examination practice. Thus the fourth sentence of § 121 provides that failure to impose a restriction requirement is not grounds of patent invalidity:
§ 121[4] The validity of a patent shall not be questioned for failure of the Commissioner to require the application to be restricted to one invention.
Section 121, viewed overall, assures that the technicalities of restriction practice are not elevated from their purpose of examination convenience to a potential taint on the validity of the ensuing patents. It was explained in the House and Senate Reports accompanying the 1952 Patent Act that § 121 means that
*1569neither of the resulting patents can be held invalid over the other merely because of their being divided in several patents.
S.Rep. No. 1979, 82d Cong., 2d Sess. 20, reprinted in 1952 USCCAN 2894, 2413; H.R.Rep. No. 1923, 82d Cong., 2d Sess. 20 (1952). Thus it is not material whether the imposition of the restriction requirement might have been avoided had different claims been presented initially, for the purpose of § 121 is simply to safeguard patent validity from the vagaries of the restriction practice, not to change the practice.
The examiner raised no objection to the inclusion of non-epitaxial claims in the ’609 process application after restriction. We affirm the district court’s ruling that the ’609 patent is not invalid for double patenting.
Obviousness
ASM states that the prior art renders the ’609 patent invalid for obviousness, 35 U.S.C. § 103. The district court, after thorough analysis of the prior art in light of ASM’s arguments, sustained the validity of the patent. ASM appeals that ruling.
ASM relies on the combination of United States Patent 3,408,982 to Capita and a 1966 text by C.F. Powell discussing chemical vapor deposition. The Capita patent describes conducting CVD using radio frequency heating. The Powell text describes the use of both radio frequency and radiant heating of CVD chambers. The district court, after extensive testimony by both sides, found that “Powell expressly suggests the use of lamps as a radiant heat source only for small scale, low temperature applications,” referring to the conclusion of Applied Materials’ expert that “nothing in Powell contemplated substitution of heat sources at all.” Applied Mate-Hols v. ASM, 32 USPQ2d at 1875. The district court gave weight to the thinking of the art at the time of the invention, concluding that “[t]he teaching of the Powell reference is that RF [radio frequency] heating is preferred for the design of CVD reactors.” Id. at 1876.
ASM places heavy reliance on a ruling of the Court of Customs and Patent Appeals during prosecution of the ’609 application. In re McNeilly, No. 79-562 (CCPA Nov. 1, 1979) (unpublished), 612 F.2d 585 (table). The examiner had rejected the claims for obviousness based on the Capita patent in view of the Powell text, stating that this combination presented a prima facie case of obviousness. The PTO Board of Appeals affirmed, as did the CCPA. The applicant then returned to prosecution before the examiner, filed a continuation application, amended the claims, filed objective evidence of the commercial success of the radiant energy process compared with the radio frequency process, and further described the unexpected results of enhanced thermal stability, reduced crystallographic slip,4 and more uniform deposition layers. The examiner eventually allowed the claims.
When a patent has been examined and duly granted, judicial review must give due weight to the presumption of validity. 35 U.S.C. § 282 (“A patent shah be presumed valid.”) The presumption of validity is based on the presumption of administrative correctness of actions of the agency charged with examination of patentability. Interconnect Planning Corp. v. Fell, 774 F.2d 1132, 1139, 227 USPQ 543, 548 (Fed.Cir.1985) (the government agency is presumed to have done its job). It was thus incumbent upon ASM to prove invalidity by clear and convincing evidence. The presentation at trial of additional evidence that was not before the PTO does not change the presumption of validity or the standard of proof, although the burden may be more or less easily carried because of the additional evidence.
ASM argues that the district court should have accepted the holding of the CCPA in In re McNeilly and declared the ’609 patent invalid. However, in Applied Materials, Inc. v. Gemini Research Corp., 835 F.2d 279, 281, 15 USPQ2d 1816, 1818 (Fed.Cir.1988), this court pointed out that the CCPA’s decision in In re McNeilly did not give rise to an estop-pel because new and material evidence was *1570presented in the PTO during the continuation of examination after the CCPA decision. The PTO’s decision to grant the patent was reached after consideration of all the evidence including the new evidence filed on return to examination. It thus became irrelevant, at this later stage, that less than all the evidence had at an earlier stage presented a prima facie case of obviousness.
Although ASM argues that the district court did not accord the opinion of the CCPA proper weight, it was pointed out in Applied Materials v. Gemini Research that the CCPA in In re McNeilly did not hold that the ’609 invention was unpatentable for obviousness; the CCPA held that the examiner had made a prima facie ease of obviousness on the evidence that was then of record. In Gemini Research this court held that summary judgment based on In re McNeilly was improperly granted. Thus the district court in the ease at bar correctly held that there was not collateral estoppel based on In re McNeilly.
ASM also argues that the CCPA’s decision in In re McNeilly is “highly persuasive” that the invention is prima facie obvious, and that the district court should have accepted that premise and considered only whether that prima facie case was successfully rebutted. ASM argues that the objective evidence was insufficient to overcome the prima facie ease of obviousness. However, the determination of obviousness, vel non, requires that all the evidence be considered together. As explained in In re Piasecki, 745 F.2d 1468, 1472, 223 USPQ 785, 788 (Fed.Cir.1984), “[i]f rebuttal evidence of adequate weight is produced, the holding of prima facie obviousness, being but a legal inference from previously uncontradicted evidence, is dissipated.” The objective evidence of unobviousness is not evaluated for its “separate knockdown ability” against the “stonewall” of the prima facie case, In re Rinehart, 531 F.2d 1048, 189 USPQ 143 (CCPA 1976), but is considered together with all other evidence, in determining whether the invention as a whole would have been obvious to a person of ordinary skill in the field of the invention.
At trial to the district court, Applied Materials presented additional objective evidence of unobviousness including evidence that four major companies (Motorola, Western Electric, Texas Instruments, and Merck) had tried unsuccessfully to develop lamp-heated CVD reactors. There was evidence of the unsuccessful attempts of these and other companies to achieve the enhanced temperature uniformity and improved chemical deposition that were achieved using the ’609 process. Witnesses described the “skepticism,” “misgivings,” and “disbelief’ in the industry when Applied Materials achieved these superior results with a radiant heating process. The district court was required to consider this evidence along with the other evidence in determining whether, on the totality of the evidence, invalidity on the ground of obviousness had been proved by clear and convincing evidence.
ASM challenges the evidentiary value of the objective evidence of commercial success on the added ground that the claims are broader than Applied Materials’ successful commercial embodiment. However, a patentee need not show that all possible embodiments within the claims were successfully commercialized in order to rely on the success in the marketplace of the embodiment that was commercialized.
All of the facts relevant to the issue of obviousness were before the district court, including the decision of the CCPA, the prosecution history in the PTO, and the evidence adduced at trial. After examining the evidence with the assistance of expert witnesses and advocates for both sides, the district court concluded that ASM had not met its burden of proving that a person skilled in this field would have been motivated by the teachings of the prior art to change from the radio frequency state-of-the-art process to a radiant heat process. No error of fact or law has been shown in that analysis. We affirm the district court’s conclusion that the ’609 patent is not invalid on the ground of obviousness.5
*1571II
THE ’389 (LOGAR) PATENT
The ’389 patent is directed to the reduction or elimination of static electricity during the CVD process. The elimination of static charges on the substrate and throughout the chamber results in purer deposition and fewer defects in the finished product, since static electricity is one of the mechanisms whereby contamination of the deposited layer occurs. The district court found that ASM’s Epsilon One process did not literally infringe the ’389 claims in suit, but that it was an infringement by application of the doctrine of equivalents.
A typical CVD process of the prior art, as described in the specification of the ’389 patent, includes several “purge” steps in which the atmosphere in the chamber is replaced by other gases. The first purges occur after the substrate is loaded in the chamber and air is replaced with nitrogen and then hydrogen; the substrate is then heated and the circuitry etching and chemical vapor deposition steps are performed, with subsequent purges occurring whenever one atmosphere is replaced by another. Static electricity is generated during the ■ initial “cold” purge steps of the process. As explained in the ’389 specification, static charges are not a problem during subsequent purges of the chamber because after the initial steps the temperature of the chamber remains above about 180° C, the temperature above which static charges do not exist.
In the invention of the ’389 patent, static charges during the initial “cold” purges are eliminated by operating the lamps at a low level during the initial gas flow steps. Claim 1 describes the energy level as a “relatively low flux intensity,” and claim 4 describes the invention in terms of heating the substrate to between 180° C and the gas phase processing temperature:
1. In a cold purge process for preparing a reactor chamber for the gas phase formation of an epitaxial layer on a wafer positioned within the reactor chamber by communicating a flow of conditioning gas into the reactor chamber, the reactor chamber system including radiant energy lamp means for heating the chamber to effect said gas phase processing, the improvement comprising:
during said gas flow step, operating the radiant energy lamp means at a power level below that used for said gas phase processing heating to supply radiant energy at a relatively low flux intensity to substantially eliminate static-induced particle transport to the wafer.
4. An improved purge process for conditioning a cold semiconductor reactor chamber for gas phase processing of a wafer therein prior to heating the wafer for the gas phase processing, comprising flowing at least one purging gas through the chamber and, during said purging step, heating the wafer to a temperature between about 180° C. and below said gas phase processing temperature, for decreasing electrostatic attraction between the wafer and particulates.
The ’389 specification describes this initial, gaseous purge as conducted at low lamp power, about six to ten percent of the deposition level. In ASM’s Epsilon One reactor this initial purge is conducted above about 850° C. Although the Epsilon One purge temperature is still below the gas phase processing temperature, ASM’s position is that it does not conduct a “cold” purge at all, but that its initial purge is really a “hot” purge and unrelated to the removal of static-induced contamination, as the claims in suit require, because such contamination does not form under the Epsilon One conditions. Thus ASM argues that although its initial purge steps ostensibly occur within the temperature limits stated in the claims, the other claim limitations of cold purge process, cold reactor chamber, and elimination of electrostatic particles, are not met literally or equivalently.
Literal Infringement
The ’389 patent contemplates a CVD process sequence that starts with a cold reactor chamber, the chamber having been cooled in order to permit the operator to remove the *1572coated substrates and insert uncoated ones. The ’389 claims are directed to the removal of static attraction during the gaseous purge of the cooled chamber within which electrostatic contamination has formed. Although the parties dispute whether this requires treating at a “relatively low” temperature that is high enough to eliminate the static charges, or is accomplished solely by an infrared photon flux at low intensity, claim 1 appears to include both mechanisms, while claim 4 is specific to heating the wafer above about 180° C. The ’389 patent describes about 180° C as the temperature above which static electricity does not form or exist.
The ASM Epsilon One reactor process also conducts an initial purge with inert gases as the first step in the deposition cycle. However, this purge is conducted at about 850° C. Since in the Epsilon One reactor the substrates are removed and replaced mechanically, the chamber need not be cooled significantly to perform this step and does not acquire electrostatic contamination during the processing cycle. Thus although the Epsilon One process includes an initial gaseous purge below the deposition temperature, the temperature of this purge is stated to remain above 850° C. ASM argues that this can not reasonably be deemed a “cold purge process.” ASM states that the problem solved by the ’389 invention does not exist at the Epsilon One operating conditions, and that the ’389 specification makes clear that the invention relates to a process wherein the chamber is cooled to ambient temperatures, permitting electrostatic contamination to form, in order to remove it by raising the temperature to above 180° C, but still at relatively low temperatures. ASM points out that the prior art shows the initial cold purging with inert gases, and that the sole “inventive” contribution of the ’389 patent is removal of the electrostatic contamination that occurred below about 180° C. ASM states that under its operating conditions, which do not drop below 850° C during or between processing cycles, electrostatic contamination does not occur, and thus that it does not practice the invention of the ’389 patent.
Applied Materials states that “cold purge” is a word of art for the initial gas purge step, and that “cold” is understood in this art as a relative term and means a temperature below the processing temperature. Applied Materials states that the claims require only that the initial gaseous purge is conducted below the processing temperature, and that ASM does so, literally. Applied Materials states that it is the claims that define the invention, and that it is incorrect to limit the claims to any specific temperature mentioned in the specification. Thus Applied Materials states that claim 1 reads literally on the Epsilon One process.
The construction of patent claims, as well as the meaning and scope of a disputed technical term or terms of art in a patent claim, are deemed to be questions of law and are determined de novo on appeal. Markman v. Westview Instruments, Inc., 52 F.3d 967, 34 USPQ2d 1321 (Fed.Cir.1995) (en banc), aff'd, - U.S. -, 116 S.Ct. 1384, 134 L.Ed.2d 577 (1996). Although this court requires that no deference be given to the testimony of experts, id. at 983, 34 USPQ2d at 1333 (“[w]hen legal ‘experts’ offer their conflicting views of how the patent should be construed, or where the legal expert’s view of how the patent should be construed conflicts with the patent document itself, such conflict does not create a question of fact nor can the expert opinion bind the court or relieve the court of its obligation to construe the claims according to the tenor of the patent. This opinion testimony also does not change or affect the de novo appellate review standard for ascertaining the meaning of the claim language.”), we take note that extensive expert testimony was adduced at trial.
The district court first considered Applied Materials’ argument that since the phrase “cold purge” appears only in the preamble of claim 1, it does not limit the claims as applied to the accused process. The district court concluded that “cold purge” is indeed an element of the claims, and “establish[es] a limitation which the accused device must meet in order to literally inflinge the ’389 patent.” We reach the same conclusion.
Whether a preamble stating the purpose and context of the invention constitutes *1573a limitation of the claimed process is determined on the facts of each ease in light of the overall form of the claim, and the invention as described in the specification and illuminated in the prosecution history. See In re Stencel 828 F.2d 751, 754, 4 USPQ2d 1071, 1073 (Fed.Cir.1987) (the preamble is interpreted in light of the invention as a whole); Perkin-Elmer Corp. v. Computervision Corp., 732 F.2d 888, 896, 221 USPQ 669, 675 (Fed.Cir.), cert. denied, 469 U.S. 857, 105 S.Ct. 187, 83 L.Ed.2d 120 (1984) (the limitations stated in the preamble give meaning to the claim and can serve to define the invention). It is thus appropriate to determine whether the term in the preamble serves to define the invention that is claimed, or is simply a description of the prior art.
The term “cold purge” has its roots in the specification, which states that in a typical deposition process of the prior art “[b]oth the nitrogen prepurge and hydrogen purge are carried out ‘cold’, that is, without heating.” The specification explains that in the ’389 process these initial purges are carried out at low radiant energy or low thermal energy. The district court found that “cold purge process” means temperatures below 180° C, and that the ’389 invention was directed to the use of heat sufficiently high to remove electrostatic contamination in the initial purge steps, that is, heat above about 180° C, in a reactor whose operating conditions include temperatures below 180° C. “Cold purge” is interpreted in light of the problem the ’389 patent solved: the elimination of electrostatic contamination during the initial purge step. The specification states:
Diming the use of the above outlined process sequence, electrostatic attraction is operative almost exclusively during the cold purging steps (steps 1 and 2). It is substantially non-existent when the sus-ceptor/wafer is at an elevated temperature, such as for example above about 180° C.
’389 patent, column 9, lines 17-22. The specification also states that the purpose of the invention is the removal of contamination caused by this electrostatic attraction. Claims 1 and 4 include this requirement.
The district court correctly placed the term “cold purge process” in the context of the state of the art when the ’389 invention was made. This context requires construing the literal meaning of the claims as limited to the process wherein electrostatic contamination is formed and removed. The Epsilon One reactor does not meet this criterion. We affirm the court’s ruling that claim 1 is not literally infringed.
The district court had previously held, on a motion for summary judgment, that claim 4 of the ’389 patent is not literally infringed because it contains the limitation “for conditioning a cold semiconductor reactor chamber” in its preamble. Applied Materials, Inc. v. Advanced Semiconductor Materials Am., Inc., No. C-91-20061-RMW(EAI) (N.D.Cal. Jan. 28, 1993). Applied Materials does not challenge this ruling on appeal.
The Doctrine of Equivalents
The district court found that claims 1 and 4 were infringed under the doctrine of equivalents. The court found that the initial purge steps of the Epsilon One reactor performed substantially the same function as that of the ’389 initial purge and achieved the same result, in that static electricity was prevented from forming. The court found that the way the Epsilon One performed this function, by conducting the initial gaseous purge above the temperature at which static electricity can exist, was substantially the same as the way in which the ’389 invention would, for in both cases static-induced contamination was eliminated by the application of heat. Concluding that the differences between the purge processes of the ’389 invention and the Epsilon One reactor were insubstantial, the district court found infringement under the doctrine of equivalents.
Infringement may be found under the doctrine of equivalents although the claim limitations are not literally met, if the accused process is substantially the same as the patented process. Hilton Davis Chem. Co. v. Warner-Jenkinson Co., 62 F.3d 1512, 35 USPQ2d 1641 (Fed.Cir.1995) (en banc), cert. granted, — U.S.-, 116 S.Ct. 1014, 134 L.Ed.2d 95 (1996). Guidance for the determination of equivalency is provided in Graver Tank & Mfg. Co. v. Linde Air Prods. Co., 339 U.S. 605, 70 S.Ct. 854, 94 L.Ed. *15741097, 85 USPQ 328 (1950), the Court holding that equivalency may be found if the accused process performs substantially the same function as the patented process in substantially the same way to obtain the same result. The doctrine of equivalents enables the trier of fact to achieve a just result when the circumstances warrant, preventing “the unscrupulous copyist [from making] unimportant and insubstantial changes and substitutions in the patent which, though adding nothing, would be enough to take the copied matter outside the claim, and hence outside the reach of law.” Id. at 607, 70 S.Ct. at 856, 85 USPQ at 330. The Court thus recognized that the claiming process is not perfect, and that judges are no less empowered to reach a just result in patent cases than in any other cause.
ASM argues that this finding of equivalency is clearly erroneous, and that there are substantial differences between the purge processes of the ’389 patent and the Epsilon One. ASM argues that while both processes purge at temperatures at which static electricity does not form, in the Epsilon One reactor there is no static-induced contamination to remove. ASM refers to the ’389 specification, which distinguished the initial gaseous purge and removal of the static electricity that is formed at ambient temperatures, with later process steps in which the higher heat within the reactor prevents electrostatic contamination:
Since the processing steps commencing with the ramp up (step 3) are either conducted at elevated temperatures or, in the case of the end nitrogen and hydrogen purges are conducted in a reactor in which there is considerable residual heat in the susceptor and elsewhere, electrostatic attraction is substantially eliminated by thermal energy.
’389 patent, column 9, lines 22-28.
ASM stresses that the invention of the ’389 patent is directed to the removal of electrostatic contamination, and that the claims so require; and that the Epsilon One reactor operates at temperatures at which electrostatic contamination does not occur. Although patent claims are not ordinarily limited to the inventor’s purpose, when that purpose is included in the claims it serves as a limitation of the claimed invention and should be met either literally or equivalently in order to satisfy the criteria of infringement. Thus although infringement is not avoided if the accused device performs an additional function beyond what is claimed, the device must nonetheless meet the requirements stated in the claim, literally or equivalently.
In this case the “cold purge process,” understood in light of the description of the ’389 invention, requires the function of removing the electrostatic contamination that is formed at low temperatures. That function is not performed in the Epsilon One. By mechanical loading and unloading of its chambers at 850° C, the Epsilon One process is conducted without the formation of electrostatic contamination. There is no interim cooling of the Epsilon One reactor to below about 180° C, and thus electrostatic attraction does not form and is not removed as claim 1 requires. This is not an insubstantial difference, Hilton Davis, 62 F.3d at 1518, 35 USPQ2d at 1645, and precludes infringement of claim 1 under the doctrine of equivalents.
Applied Materials points out that claim 4 does not contain the “cold purge process” limitation, but describes the initial purge as for “conditioning a cold semiconductor reactor chamber.” However, as the district court correctly determined, a cold reaction chamber, as that term is used in the ’389 patent, is one in which static electricity can form, ie., one whose temperature is below about 180° C. The Epsilon One reactor is heated at about 850° C while new substrates are loaded. Claim 4, like claim 1, expressly requires “decreasing electrostatic attraction” during the purging step. We conclude that the differences between the Epsilon One process and the process of claim 4 are not insubstantial, for they do not perform this function literally or equivalently. We conclude that the district court’s finding of infringement by equivalency is clearly in error.
The ruling of infringement of the ’389 patent in terms of the doctrine of equivalents is reversed.
*1575III
THE ’313 PATENT
Chief Judge ARCHER and Circuit Judge MAYER concur in affirming the judgment of invalidity of the ’313 patent, thus the judgment is affirmed.
Circuit Judge NEWMAN dissents.
COSTS
No costs.
AFFIRMED IN PART, REVERSED IN PART, AND REMANDED MOTION DISMISSED
. Applied Materials, Inc. v. Advanced Semiconductor Materials America, Inc., 32 USPQ2d 1865 (N.D.Cal.1994).
. 1. In a process for effecting chemical vapor deposition coating of a substrate in a reaction chamber, the steps of:
placing the substrate in the reaction chamber on a susceptor of substantially greater size than the substrate and adapted to absorb and conduct radiant heat energy,
transmitting radiant heat energy through a wall of the reaction chamber to the susceptor to heat the susceptor by absorption of the radiant heat energy without appreciable absorption of said energy by the chamber wall,
maintaining the substrate in contact with the susceptor to permit a transfer of heat between the susceptor and the substrate whereby the substrate is heated and maintained at a substantially uniform temperature throughout,
and introducing a gaseous reactant into the reaction chamber for contact with the heated substrate to form a film by chemical vapor deposition thereon.
. An epitaxial layer is one that has the same crystallographic structure as the substrate upon which it is deposited.
. Crystallographic slip occurs when thermal stresses within the substrate or coating cause adjacent layers of molecules to move relative to each other, thus disrupting the pattern of the deposition.
. Applied Materials has moved this court to take judicial notice of an assertedly inconsistent position of ASM in related litigation concerning the ’609 patent. In view of our decision uphold*1571ing the validity of the '609 patent on the grounds presented, the question of judicial estoppel need not be considered in this case. The motion is dismissed as moot.