These are cross-appeals from the December 16, 1985, judgment of the United States District Court for the Northern District of California holding both of Coherent's patents in suit, No. 4,378,600 entitled "Gas Laser" issued on March 29, 1983, to James L. Hobart (the Hobart patent) and No. 4,376,328 entitled "Method of Constructing Gaseous Laser" issued on March 15, 1983, to Wayne S. Mefferd (the Mefferd patent), invalid for lack of enabling disclosure under 35 U.S.C. Sec. 112, first paragraph, after originally entering judgment on a jury verdict finding claims 2, 5, 7, and 18 of the Hobart patent and claim 10 of the Mefferd patent valid and infringed by Spectra-Physics, Inc. (Spectra).

We reverse the district court's holding that both patents are invalid for lack of enablement. We also reverse, however, the court's finding that both patent specifications complied with the best mode requirement of Sec. 112, and thus affirm the judgment that the patents are invalid, but on a different legal ground.

Before discussing the legal aspects of this case, we first explain the technology involved which gave rise to them.

The Hobart patent is directed to an ion laser structure and the Mefferd patent to a method of fabricating an ion laser. "Laser" is an acronym for l ight a mplification by § timulated e mission of r adiation.1 An ion laser is a type of gaseous laser. The lasing medium, typically argon or krypton gas, is contained within a sealed discharge tube which is axially aligned with a pair of mirrors to form the optical cavity or resonator.

For lasing to take place, the argon or krypton gas must be excited to elevated energy states. This is accomplished by providing a high-energy electrical discharge through the gas. The discharge through the laser must then be constrained to a straight line along the laser's optical path and pinched to a small diameter to concentrate its energy into a small elongated volume.

The discharge through the laser is extremely hot--up to 6000 degrees C. The exterior of the laser, however, must operate at room temperature, requiring dissipation of large amounts of heat by external cooling. It is also important that gas pressure be uniformly controlled along the discharge tube. For some reason, not entirely agreed upon by physicists, the gas tends to move to one end of the tube or the other. This phenomenon, known as "pumping," causes an uneven gas pressure differential in the discharge tube, resulting in poor performance or no performance at all.

The Hobart patent is directed to a gas laser having an improved laser discharge tube.2 The discharge path of the laser is determined by coaxially aligned apertures in a series of spaced-apart tungsten discs within the laser discharge tube. The discharge tube itself is a thin-walled ceramic tube, for example, of alumina (Al2 O3 ). Heat from the tungsten discs is transmitted by conduction to and through the ceramic tube (26), which is surrounded by a water jacket, by copper cups (50) attached to the inside of the tube. See Fig. 1 below, which is a dissected sectional view showing two end portions with a substantial portion of the central section omitted, the broken line representing the longitudinal axis.

NOTE: OPINION CONTAINS TABLE OR OTHER DATA THAT IS NOT VIEWABLE

Figure 1.

Cross-sectional View of Laser StructureClaim 1 calls for "means for providing a heat conduction path from the central aperture of each of said discs to and through the tube wall." This includes both means for attaching the tungsten discs (48) to the center opening of each copper cup and means for attaching the cups to the inside wall of the ceramic tube. High thermal conductivity is achieved by brazing or soldering which provides a permanent metallic contact between the cups and the tube wall.

The Hobart patent further discloses and claims a "shield" feature which is a cylindrical ring coaxially attached to or formed integrally with each of the copper cups.3 These shields (56) aid in minimizing gas pumping within the discharge tube.

The Mefferd patent describes a method of fabricating the laser structure of the Hobart patent.4 The problem addressed in the Mefferd patent is how to insert and hold in place the heat conducting cups inside the long, slender tube, while at the same time maintaining the apertured discs in precise alignment. The patent discloses a "floating" disc technique whereby the disc apertures are aligned by tensioning a mandrel that has been passed through each of the disc apertures. Once the disc apertures are aligned, the whole assembly is brazed to permanently bond the parts within and to the tube walls. See Fig. 12 below in which the copper cups are 50, the shields 56, the discs 48, the mandrel 74 and the ceramic tube 26. The figure shows a partial assembly before the brazing of the discs to the cups, which is done in a vertical position with end "B" upward.

NOTE: OPINION CONTAINS TABLE OR OTHER DATA THAT IS NOT VIEWABLE

Figure 12.

Cross-sectional View Illustrating Assembly of Cups and Discs

Both the Hobart and Mefferd patents stress the importance of the bond between the copper cups and the ceramic tube. Poor thermal contact between them results in higher disc temperatures which in turn impedes the gas flow through the tube. For the laser to be reliable, the copper-ceramic bond must also be able to withstand repeated heat cycling. Due to the differing rates of thermal expansion of copper and alumina, the bond is subject to compressive forces as the laser heats up and tensile stress during cooling.

Dr. Hobart initially approached the problem of how to make the critical copper to ceramic bond by experimenting with soldering. These attempts were unsuccessful and no attempt was made to even try to solder together any laser shaped parts. Wayne Mefferd was then brought in to solve the attachment problem. His solution was brazing.

While the patent specifications disclose pulse soldering as one method of attachment, brazing is clearly the preferred method. In this process, a brazing shim 68, Fig. 4, is placed between the copper cup 50 and the inner wall of the ceramic tube 26, see Fig. 3, and the whole assembly is heated to the melting point of the braze material.

NOTE: OPINION CONTAINS TABLE OR OTHER DATA THAT IS NOT VIEWABLEDuring heating the cup is mechanically expanded into contact with the tube by means of an expansion tool inserted into tube 26.

The patents further disclose "TiCuSil" as the preferred brazing material. This material is a copper silver eutectic (an alloy whose ingredients are proportioned to have the lowest possible melting point) with a small percentage of titanium added for making a ceramic to metal seal under what is known as the active metal process. In this process, the titanium invades and wets the ceramic so that the copper-silver braze material can hold the copper to the ceramic. In the absence of an active metal alloy component such as titanium, the ceramic must be premetalized with, for example, moly-manganese (MoMn), to provide a metallic surface to which the copper-silver braze material will adhere.

The TiCuSil active metal process is preferred because it requires only one step and avoids the need for premetalization. In addition, the copper cups cannot be electrically connected because this destroys the evenly graduated electrical potential down the bore of the tube which is required for the laser to operate. Thus, any premetalization must be in circular stripes along the inner surface of the tube so that each copper cup can be brazed or soldered to a different stripe.

E. Patentee Coherent's Six-Stage Braze Cycle

According to the standard product specification sheet, TiCuSil should be brazed at 850 degrees C. The sheet also specifies that the braze should be performed in a vacuum or in a neutral atmosphere of dry argon gas. Using these general guidelines, Mefferd developed a six-stage braze cycle for using TiCuSil to attach the copper cups to the ceramic tube. "Braze cycle" is a term of art which refers to a process defined by specific parameters of temperature, length of times at given temperatures, atmosphere, and pressure.

Mefferd knew that there were tradeoffs in the braze cycle. For one, it is generally desirable to heat the parts as fast as possible. As the parts are heated, however, "outgassing" occurs and contaminants trapped in the parts are released into the atmosphere of the oven. The vacuum pump removes the outgassed contaminants, but if the outgassing is too rapid, then the pressure may rise and the pump will not work. Also, if oxygen is evolved as part of the out-gas, the titanium may react with it and degrade the strength of the braze joint.

In assessing the tradeoffs, Mefferd let the pressure control the braze cycle, as one experimental approach. For example, Mefferd held the pressure in the oven at 10-4 torr while the assembly was initially heated from 0 to 750 degrees C. This took from an hour and a half to two hours. In the next step, he held the temperature at 750 degrees for ten to fifteen minutes while further reducing the pressure. The temperature was then increased for a period of approximately fifteen minutes until it reached 850 degrees C. The remaining steps include holding the temperature at 850 degrees C for a specified time and at a pressure of 10-4 torr, decreasing the oven temperature to 750 degrees C and adding argon gas, and then turning the oven off.

Mefferd's six-stage cycle produced a reliable braze joint between the copper cups and the ceramic tube. Because this approach worked, Coherent continued to use TiCuSil and never investigated the moly-manganese process or further experimented with soldering. Neither the Hobart patent nor the Mefferd patent, however, discloses the braze cycle or any additional information on brazing copper to ceramic using TiCuSil.

Dave Wright, head of research at Spectra, and his technician Martin Riley, worked on so-called "cold disc" lasers of the type in suit in the late 1970's. They referred to these lasers as "cold disc" lasers because the process of brazing the copper cups to the ceramic tube provided good thermal conduction as contrasted with the earlier radiatively cooled lasers which ran hot. Wright and Riley, however, had only limited success with cold disc lasers, in part because they could not make a satisfactory bond between the copper cups and the alumina ceramic tube. Upon repeated heat cycling, the ceramic would crack and cause the copper to break away, overheat, and melt, which destroyed the operation of the tube. After two and a half years, Wright was unable to make a TiCuSil braze joint which was reliable enough for a commercially acceptable product and Spectra temporarily abandoned the project.

Spectra resumed work on the cold disc project in 1981 after Coherent introduced its INNOVA laser embodying the inventions of the patents in suit. Because of their uncertainty about brazing, Spectra hired a brazing expert, Dr. Leonard Reed, to develop a moly-manganese process for attaching the cups to the ceramic tube. After nearly a year of experimentation, Dr. Reed developed Spectra's proprietary moly-manganese process. This involved using precision ceramic tubing and a special computerized striping tool which ground circular rings away from a coat of moly-manganese metallization painted on the entire inside of the tube.

Spectra eventually introduced its Model 2020 gas laser which was made using the moly-manganese process. Like Coherent's laser, each of the copper cups in the 2020 laser has a ring or shield to alleviate the gas pumping problem.

Spectra brought a declaratory judgment action against Coherent asking a holding of invalidity and non-infringement of both patents. Coherent counterclaimed for infringement and an adjudication of validity.

After thirty-two days of trial, the district court submitted eighteen "Interrogatories to the Jury" and the jury answered fifteen of them, finding in part that Spectra's model 2020 laser infringed claims 2, 5, 7, and 18 of the Hobart patent and claim 10 of the Mefferd patent (the shield claims). The jury found that the remaining claims of both patents were invalid for obviousness or were anticipated by the work done by Wright and Riley at Spectra. The jury also found, however, that Wright and Riley had not reduced their cold disc laser to practice before conception of the subject matter set forth in the Hobart and Mefferd patents.

After entering judgment on the jury verdict, the district court withdrew the judgment and asked the parties to prepare proposed findings of fact and conclusions of law on several additional issues including disclosure of best mode and enablement under Sec. 112, co-inventorship of the shield claims by Dr. Rempel, another Coherent employee, and inequitable conduct. The court ruled in favor of Coherent on all of these issues except enablement. On that issue, the court held both patents invalid for failure to disclose the six-stage braze cycle used by Coherent to manufacture the laser. The court found that the best mode requirement was satisfied, however, because neither Hobart nor Mefferd deliberately or accidentally concealed brazing as the best mode of attaching the copper cups to the ceramic tube.

Coherent appeals from the judgment with respect to lack of enablement and seeks reinstatement of the jury verdict that the shield claims are valid and infringed by Spectra. Coherent also appeals the portion of the judgment finding the remaining (non-shield) claims invalid for obviousness because of an erroneous jury instruction and requests a remand for a new trial on these claims.

Spectra cross-appeals from the judgment as it relates to best mode, inventorship of the shield claims, and derivation of the claimed subject matter from Wright and Riley. Spectra also appeals the denial of its request for attorney fees.

1. Introduction--Adequate Disclosure Under Sec. 112, 1st

To constitute adequate disclosure under the first paragraph of 35 U.S.C. Sec. 112, a patent specification must set forth both the manner and process of making and using the invention (the enablement requirement) and the best mode contemplated by the inventor of carrying out the invention (the best mode requirement). The difference between these two is explained in In re Gay, 309 F.2d 769, 135 U.S.P.Q. (BNA) 311 (CCPA 1962):

The essence of [the enablement requirement] is that a specification shall disclose an invention in such a manner as will enable one skilled in the art to make and utilize it. Separate and distinct from [enablement] is [the best mode requirement], the essence of which requires the inventor to disclose the best mode contemplated by him, as of the time he executes the application, of carrying out his invention. Manifestly, the sole purpose of this latter requirement is to restrain inventors from applying for patents while at the same time concealing from the public preferred embodiments of their inventions which they have in fact conceived.

... The question of whether an inventor has or has not disclosed what he feels is his best mode is, however, a question separate and distinct from the question of the sufficiency of his disclosure to satisfy the requirements of [enablement].

Id. at 772, 135 USPQ at 315 (emphasis in original).

Thus, compliance with the best mode requirement focuses on a different matter than does compliance with the enablement requirement. Enablement looks to placing the subject matter of the claims generally in the possession of the public. If, however, the applicant develops specific instrumentalities or techniques which are recognized at the time of filing as the best way of carrying out the invention, then the best mode requirement imposes an obligation to disclose that information to the public as well. See Flick-Reedy Corp. v. Hydro-Line Mfg. Co., 351 F.2d 546, 550-51, 146 U.S.P.Q. (BNA) 694, 697 (7th Cir.1965), cert. denied, 383 U.S. 958, 86 S. Ct. 1222, 16 L. Ed. 2d 301 (1966); Union Carbide Corp. v. Borg Warner Corp., 550 F.2d 355, 361-63, 193 U.S.P.Q. (BNA) 1, 6-7 (6th Cir.1977).

The situation before us is one in which the patent specifications disclose more than one means for making the claimed invention, but do not adequately disclose the best means actually known to the inventors. The district court recognized that the specifications were inadequate under Sec. 112, but incorrectly based its decision on a lack of enablement. As we explain, the problem is really one of best mode, and thus, while we disagree with the district court's views on these issues, the judgment that the patents are both invalid was correct and must be sustained.

a. The Jury Question

Before addressing the substance of the district court's decision on enablement, we consider the question, raised by Coherent, whether the court was free to decide enablement at all without first considering the jury verdict. While no specific question was submitted to the jury on enablement, the district court instructed the jury that "invalid claims cannot be infringed" and provided general instructions on the law of enablement. Coherent maintains that by finding the shield claims infringed, the jury implicitly decided enablement in its favor, and that the court could not overrule these findings without making the determinations required by JNOV.

The district court, however, did not feel constrained by the jury verdict because it considered the question of enablement to be one of law. The court also viewed the form of the verdict as a special verdict under Rule 49(a), Fed.R.Civ.P., and not as a general verdict.

Although enablement is ultimately a question of law, see, e.g., Moleculon Research Corp. v. CBS, Inc., 793 F.2d 1261, 1268, 229 U.S.P.Q. (BNA) 805, 810 (Fed.Cir.1986), cert. denied, --- U.S. ----, 107 S. Ct. 875, 93 L. Ed. 2d 829 (1987), this court has recognized that there may be underlying factual issues involved, see Quaker City Gear Works, Inc. v. Skil Corp., 747 F.2d 1446, 1453-54, 223 U.S.P.Q. (BNA) 1161, 1166 (Fed.Cir.1984), cert. denied, 471 U.S. 1136, 105 S. Ct. 2676, 86 L. Ed. 2d 694 (1985). The court may submit legal issues such as enablement to the jury under Rule 49(a), but if it does, the court may not make subsequent findings which overrule an implicit and inherent finding of the jury. Id. Because the district court erred in applying the substantive law of enablement, however, we need not decide whether it also overruled the jury's finding.

b. Enablement By Alternative Means

To be enabling under Sec. 112, a patent specification must disclose sufficient information to enable those skilled in the art to make and use the claimed invention. See, e.g., Hybritech Inc. v. Monoclonal Antibodies, Inc., 802 F.2d 1367, 1384, 231 U.S.P.Q. (BNA) 81, 94 (Fed.Cir.1986), cert. denied, --- U.S. ----, 107 S. Ct. 1606, 94 L. Ed. 2d 792 (1987). The district court held both of the patents in suit invalid for lack of enablement based on their failure to disclose Coherent's six-stage braze cycle for brazing TiCuSil. The court found that the braze cycle was "necessary to the enjoyment of the invention [sic]."

Coherent's braze cycle, however, is applicable only to TiCuSil brazing, which is just one of the ways to make and use the claimed inventions. The Hobart patent calls for "means for attaching" the copper cups to the inside of the ceramic tube and Mefferd has essentially the same step of "permanently securing" the cups to the tube. The specifications identify as suitable attachment techniques the alternatives of TiCuSil brazing, moly-manganese brazing, and low-temperature pulse-soldering.

If an invention pertains to an art where the results are predictable, e.g., mechanical as opposed to chemical arts, a broad claim can be enabled by disclosure of a single embodiment, In re Cook, 439 F.2d 730, 735, 169 U.S.P.Q. (BNA) 298, 301 (CCPA 1971); In re Vickers, 141 F.2d 522, 527, 61 U.S.P.Q. (BNA) 122, 127 (CCPA 1944), and is not invalid for lack of enablement simply because it reads on another embodiment of the invention which is inadequately disclosed, see Gould v. Mossinghoff, 711 F.2d 396, 400, 219 U.S.P.Q. (BNA) 393, 396 (D.C.Cir.1983).5 Thus, it is sufficient here with respect to enablement that the patents disclose at least one attachment means which would enable a person of ordinary skill in the art to make and use the claimed inventions. Because the patents disclose the alternatives of moly-manganese brazing and pulse-soldering, their failure to also disclose Coherent's TiCuSil braze cycle is not fatal to enablement under Sec. 112.

Spectra argues that the patents' references to the "moly-manganese process" is only in regard to low-temperature pulse-soldering, not brazing. We disagree. A fair reading of that paragraph as part of the general discussion of brazing, given that moly-manganese brazing was the most common method of bonding metal to ceramic, is that one skilled in the art would recognize that moly-manganese brazing was an alternative means of attachment. Spectra's Dave Wright, among others, testified that moly-manganese brazing was common in the industry and was well-known for brazing copper to ceramic.

The district court ignored the moly-manganese process, however, for the erroneous reason that it was "neither described nor advocated in the patents in suit." A patent need not teach, and preferably omits, what is well known in the art. Hybritech, 802 F.2d at 1384, 231 USPQ at 94. While there is no elaboration of moly-manganese brazing in the patent specifications, the district court found that brazing was an old and well-known technique when the applications were filed.

Spectra argues that moly-manganese brazing suitable for use in constructing the lasers of the two patents was not enabled because it required undue experimentation as evidenced by the amount of time and money it spent developing its moly-manganese process. Spectra's efforts, however, were not simply directed to finding a moly-manganese process that would work, but encompassed a whole range of enterprises necessary to making a commercially successful product. In fact, it took only three months from the time Spectra decided on a specific design for its laser until it established a workable braze technique, but it was almost another year before it made its first truly successful prototype laser.

The two major problems that Spectra claims it had to solve in order to have a successful moly-manganese process also do not show lack of enablement. For example, Spectra contends that the moly-manganese process requires the use of precision ceramic tubing, something not taught in the patent. But Steve Jarrett, charged with developing Spectra's braze process, stated that the reason he used precision tubing was to save labor costs. Likewise, Spectra claims that Dr. Reed had to develop a special computerized striping tool to grind away circular rings from the coat of moly-manganese metallization painted on the inside of the tube. A Spectra in-house report reveals, however, that stripe metallized tubes were available from the same vendors that made the ceramic tubes themselves.

As for the court's statement that moly-manganese was not "advocated" in the patents, this is another matter entirely. We can only surmise that the court somehow confused the enablement requirement with the best mode.6 Nonenablement is the failure to disclose any mode, In re Glass, 492 F.2d 1228, 1233, 181 U.S.P.Q. (BNA) 31, 35 (CCPA 1974), and does not depend on the applicant advocating a particular embodiment or method for making the invention. In practical terms, where only an alternative embodiment is enabled, the disclosure of the best mode may be inadequate. But that is a question separate and distinct from the question whether the specification enabled one to make the invention at all. In re Gay, 309 F.2d at 772, 135 U.S.P.Q. (BNA) at 315.

Finally, there is no mention at all of low-temperature pulse-soldering, except to the extent that the court considered it to be the same as the "moly-manganese process." Spectra asserts that pulse-soldering could not be used in the method claimed in the Mefferd patent because it permits soldering of only one cup at a time. This may be true, but it says nothing about making the structure claimed in the Hobart patent. At the very least, the court should have considered whether it was an operative alternative for making the Hobart structure.

a. The Jury Question (Again)

Coherent raises the same question with respect to best mode that we found unnecessary to decide for enablement, that is, whether the jury implicitly found that the patents disclosed the best mode. Because the district court also found for Coherent on best mode, it would seem unnecessary to decide the jury question in this context as well. It is relevant, however, as far as it affects the proper standard of review in this court and so we must address it.

The parties clearly intended the jury to decide best mode. The interrogatories submitted to the jury, however, failed to carry out that intent. Interrogatory 14 read as follows:

14. Did Coherent disclose the best mode known to it at the time the patent application was filed?

Yes ---------- No ----------

The question came back unanswered because the preamble to the question instructed the jury to consider best mode only if it had answered the preceding interrogatory on obviousness "no." The jury answered that question "yes"--but only for certain claims. The preamble to interrogatory 14 should have instructed the jury to answer best mode if the answer to obviousness was "no" or "yes as to less than all claims."

This is the type of inadvertent omission that Rule 49(a), Fed.R.Civ.P., remedies by providing that trial by jury has been effectively waived in these circumstances. See Quaker City Gear Works, Inc. v. Skil Corp., 747 F.2d 1446, 1453, 223 U.S.P.Q. (BNA) 1161, 1165 (Fed.Cir.1984), cert. denied, 471 U.S. 1136, 105 S. Ct. 2676, 86 L. Ed. 2d 694 (1985) (right to trial by jury of factual issue may be waived by agreeing to instructions that jury need not answer all questions); see also 5A Moore's Federal Practice p 49.03 (1986). Coherent's failure to object to the form of the interrogatories caused the best mode issue, by operation of Rule 49(a), to revert to the court for decision. See Quaker City, 747 F.2d at 1453, 223 USPQ at 1166. Thus, we review the district court's findings on best mode under the "clearly erroneous" standard of Rule 52(a), Fed.R.Civ.P., and not, as Coherent's argument suggests, as if the lower court's ruling was a denial of a motion for JNOV under Rule 50(b).

The single instruction to the jury that invalid claims cannot be infringed (a nonsense statement), one of many on supposed general principles of patent law, does not operate to convert the interrogatories on infringement into general verdicts which subsumed all of Spectra's invalidity defenses, including best mode. Any inference that the jury implicitly found that the best mode requirement was satisfied is negated by the preamble to interrogatory 14 which in effect told the jury that it need not consider the best mode.

b. Adequate Disclosure of Best Mode

Because the best mode provision of Sec. 112 speaks in terms of the best mode "contemplated by the inventor," there is no objective standard by which to judge the adequacy of a best mode disclosure. DeGeorge v. Bernier, 768 F.2d 1318, 1324, 226 U.S.P.Q. (BNA) 758, 763 (Fed.Cir.1985). Instead, only evidence of "concealment," whether accidental or intentional, is considered. Id. The specificity of disclosure required to comply with the best mode requirement must be determined by the knowledge of facts within the possession of the inventor at the time of filing the application. See United States Dep't of Energy v. Daugherty, 687 F.2d 438, 446, 215 U.S.P.Q. (BNA) 4, 11 (CCPA 1982).

Compliance with the best mode requirement, because it depends on the applicant's state of mind, is a question of fact subject to the clearly erroneous standard of review. See McGill, Inc. v. John Zink Co., 736 F.2d 666, 676, 221 U.S.P.Q. (BNA) 944, 951 (Fed.Cir.), cert. denied, 469 U.S. 1037, 105 S. Ct. 514, 83 L. Ed. 2d 404 (1984). This assumes, however, a proper legal understanding of the best mode requirement, which we find missing from the district court's analysis. In general, we do not disagree with the facts as found by the district court. It is only the court's ultimate conclusion that the best mode requirement was satisfied that we reject.

In analyzing compliance with the best mode requirement, the district court focused only on the generic rather than the specific information known to the inventors and found that neither Mefferd nor Hobart intentionally, deliberately, or accidentally "concealed the braze technique as the best mode of attaching the heat web to the alumina tube." (Findings of Fact 4 and 8.) The patent specifications make clear, however, that the best mode contemplated by the inventors, as least as far as the critical "means for attaching" the copper cups to the ceramic tube is concerned, was more than just brazing in general--it was TiCuSil active metal brazing. Coherent acknowledges as much by its references to TiCuSil as the "preferred" brazing material and by the fact that Coherent never used anything else.

The appropriate question then is not whether the inventors disclosed TiCuSil brazing at all--they did--but whether TiCuSil brazing was adequately disclosed. See In re Sherwood, 613 F.2d 809, 816, 204 U.S.P.Q. (BNA) 537, 544 (CCPA 1980), cert. denied, 450 U.S. 994, 101 S. Ct. 1694, 68 L. Ed. 2d 193 (1981). Even though there may be a general reference to the best mode, the quality of the disclosure may be so poor as to effectively result in concealment. Id.

The facts found by the district court, when placed in the proper framework, plainly demonstrate that the TiCuSil brazing technique used by Coherent was not adequately disclosed. The court stated in findings of fact under the heading "ENABLEMENT":

2. The use to which Coherent put the TiCuSil braze material was, and was known to be by Coherent at the time, contrary to criteria for the use of TiCuSil as contained in the literature.

and again,

9. The references to brazing as used in the patents and the extraneous texts (Kohl, Wesgo Brochure) relied upon by Coherent, do not describe for the benefit of one skilled in the art of laser construction the manner in which the Mefferd method is usable for the construction of the Hobart apparatus by means of [TiCuSil] brazing.

The district court also found that the inventors were aware of the problems associated with TiCuSil:

4. The known difficulty recognized by Hobart and Mefferd in working with TiCuSil as a braze material for the purpose to which they put it is reflected in Hobart's disclosure dated March 1, 1979 ... that the titanium-copper-silver process is "not in high favor in the ceramic industry" and "not preferred as compared with what is called the moly-manganese technique which produces stronger and also less leak-prone seals" and essentially the same language in the May, 1979 patent disclosure signed by all of Hobart, Mefferd and Johnston.

Coherent admits that its braze cycle is not disclosed in either patent nor is it contained in the prior art. Instead, it maintains that its braze cycle is unique to its ovens, and because the performance of industrial ovens varies considerably, the actual parameters would be meaningless to someone who used a different oven. In support of its position, Coherent cites In re Gay, 309 F.2d at 769, 135 USPQ at 316, which states that "[n]ot every last detail is to be described, else patent specifications would turn into production specifications, which they were never meant to be." In doing so, however, Coherent was not discussing whether it had complied with the best mode requirement because the court had held in its favor on that issue; it was discussing whether it had complied with the enablement requirement on which the court had held against it.

First, it is not up to the courts to decide how an inventor should disclose the best mode, but whether he has done so adequately under the statute. Weil v. Fritz, 601 F.2d 551, 555, 202 U.S.P.Q. (BNA) 447, 450 (CCPA 1979). Second, far from being a "production specification," Coherent did not disclose any details about its brazing process. It is this complete lack of detail which effectively resulted in its concealment.

Where the district court went wrong on the law while reaching the right result is starkly revealed in its conclusions of law. Under the heading of "BEST MODE" is this conclusion:

4. There was no concealment deliberate or otherwise by Hobart or Mefferd of the brazing process as the best mode of bonding the heating web to the alumina tube.

As we have pointed out, however, this refers to brazing in general, not the actual brazing cycle with TiCuSil and all of the parameters which Coherent found to be its best mode, admittedly not disclosed. In contrast, but under the heading "ENABLEMENT," is the key conclusion of law which supports our conclusion and the judgment, reading as follows:

3. The six stage braze cycle employed by Coherent, and developed by it, are [sic, is] necessary to the enjoyment of the invention taught by the patents in suit by a person skilled in the art of laser construction, and are [sic] not sufficiently disclosed by the patents in suit. [Original emphasis.]

For reasons above explained, Coherent's failure to disclose its "six stage braze cycle" fully supports the defense of non-compliance with the best mode requirement of the first paragraph of Sec. 112 although the inventions as broadly claimed could be practiced without knowledge of it, which means that the patent specifications are enabling. The trial court evidently had a grasp on the essential facts but somehow got them into the wrong legal pigeonholes. With the aid of lawyers, this is not difficult to do.

Spectra's claim in this declaratory judgment complaint that the two patents in suit are invalid must therefore be sustained on the ground that they fail to disclose the best mode contemplated by the inventors for practicing their respective inventions. 35 U.S.C. Secs. 112 and 282.

The primary basis for Spectra's request for attorney fees is Coherent's supposedly bad faith conduct during litigation, specifically the manipulation and suppression of evidence. While bad faith conduct during litigation may make a case exceptional under 35 U.S.C. Sec. 285, see, e.g., Standard Oil Co. v. American Cyanamid Co., 774 F.2d 448, 455, 227 U.S.P.Q. (BNA) 293, 298 (Fed.Cir.1985), judgments of a district court concerning good and bad faith are not easily overturned, Western Marine Electronics v. Furuno Elec. Co., 764 F.2d 840, 847, 226 U.S.P.Q. (BNA) 334, 339 (Fed.Cir.1985).

Spectra lists alleged abuses by Coherent, all presented to the court below and found unconvincing. Spectra has not shown that the court's implicit determination that this was not an exceptional case was erroneous or that the court abused its discretion by denying Spectra its attorney fees. See Reactive Metals Alloys Corp. v. ESM, Inc., 769 F.2d 1578, 1582-83, 226 U.S.P.Q. (BNA) 821, 824 (Fed.Cir.1985). Furthermore, we see no need to remand as in S.C. Johnson & Sons, Inc. v. Carter-Wallace, Inc., 781 F.2d 198, 201, 228 U.S.P.Q. (BNA) 367, 369 (Fed.Cir.1986), where the court's reasons can be readily inferred from the record and arguments made below.

Because we affirm the judgment that the Hobart and Mefferd patents are invalid, we need not decide the other issues presented by Coherent's appeal and Spectra's cross-appeal, namely, those relating to the jury instruction on obviousness, inventorship of the shield claims, and derivation from Wright and Riley.

The judgment of the district court that the Hobart and Mefferd patents are both invalid is affirmed but on a different ground than that relied on by the court below. We hold that both patents are invalid under Sec. 112, first paragraph, for failure to disclose the best mode, not for lack of enablement as the district court held.

The district court's denial of Spectra's request for attorney fees is affirmed.

ARCHER, Circuit Judge, concurring.

I join the opinion of the majority, except that I find no basis for the majority's comment regarding the "aid of [the] lawyers." The record on appeal does not indicate that the lawyers misled the court or otherwise affirmatively contributed to the court's error regarding enablement and best mode which, whether or not intended, seems to be the implication of the comment.