In re Boesch , 617 F.2d 272 ( 1980 )

MILLER, Judge.

This is an appeal from a decision of the Patent and Trademark Office (“PTO”) Board of Appeals (“board”) which sustained the examiner’s rejection under 35 U.S.C. § 103 of appellants’ claims ¹ 1 and 8-15 in view of Lamb² and Pohlman³ et al. We affirm.

Invention

The invention embraces nickel base alloys consisting essentially of:

Metals Percentage Ranges

aluminum 4.0 - 4.7

boron 0.005 - 0.03

carbon 0.0 - 0.18

chromium 13.7 - 15.3

cobalt 14.2 - 19.0

iron 0.0 . - 4.0

molybdenum 3.8 - 4.8

titanium 3.0 - 3.7

The remainder of the alloys comprises nickel and incidental impurities. The elements in the alloys are balanced to provide an Nv⁴ value not in excess of about 2.35⁵ according to the following equation:

Nv = 4.66 (A% Cr + A% Mo) + 1.71 (A% Co) + 0.61 (A% Ni)⁶

In the case of alloys within the broad range set forth above, but not balanced to meet the required Nv value, room temperature ductility deteriorates, and creep⁷ deformation increases, after prolonged exposure at elevated temperatures. Appellants state that these results are attributable to formation of a deleterious phase (known as “sigma phase”) in the metal after such exposure, and that the tendency of an alloy to form sigma phase is (unexpectedly) eliminated by balancing the relative amounts of its constituent elements in accordance with the Nv equation. Where the composition of an alloy has been controlled to provide an Nv value of about 2.35 or less, no sigma has been found after exposure at 1500 0 F for time periods up to 7200 hours.

Claim 1 is illustrative:

1. A nickel base alloy having a composition consisting essentially of up to 0.18% carbon from about 14.2% to about 19.0% cobalt, from about 13.7% to about 15.3% chromium, from about 3.8% to about 4.8% molybdenum, from about 3.0% to about 3.7% titanium, from about 4.0% to about 4.7% aluminum, up to about 4.0% iron, from 0.005% to about 0.03% boron and the balance essentially nickel with incidental impurities, the aforementioned elements being balanced to provide an Nv value not in excess of about 2.35 according to the following equation:

Nv = 4.66 (A% Cr + A% Mo) + 1.71 (A% Co) + 0.61 (A% Ni)

the alloy being characterized by its freedom from precipitation of deleterious amounts of sigma-like phase after exposure at temperatures in excess of 1500 0 F for periods of time in excess of 1000 hours.

Prior Art

Lamb discloses a process for hot working age-hardenable nickel-chromium alloys. The alloys contain:

Metals Percent by Weight

aluminum 4.0 - 5.4

boron 0.003 - 0.1

chromium 14.0 - 16.0

carbon 0.01 - 0.2

cobalt 14.0 - 25.0

molybdenum 3.0 - 5.5

titanium 3.0 - 4.6

zirconium 0.01 - 0.2

A sample alloy is heated at 1190 ° C for 1.5 hours and cooled to 1000 ° C at about 1 ° C per minute, after which it may be hot worked at 1120 ° C. When hot working is complete, the alloy will generally require a further heat treatment to develop full creep resisting properties.

Pohlman et al. disclose nickel base alloys suitable for elevated temperature operation containing:

Metals Percent by Weight

aluminum 4.2 - 4.6

boron 0.025 - 0.035

carbon 0.04 - 0.07

chromium 14.5 - 15.5

cobalt 14.5 - 15.5

molybdenum 4.5 - 5.5

titanium 3.3 - 3.7

The remainder of the alloys essentially comprises nickel and incidental impurities; possibly, also, small amounts of silicon and manganese.

Both references are silent regarding an ■ Nv value requirement, although Lamb requires “a total aluminum and titanium content from about 7.75% to about 9.5%,” and Pohlman et al. “prefer about 14.5-15.5 percent by weight cobalt because that range results in the best balance at elevated temperatures between such properties as tensile and rupture strengths, oxidation resistance and the ability of the sheet material to be formed or worked.”

The Boesch Affidavit

Seven heats of alloys (appellants’ Table I below), which were within the claimed composition ranges but whose Nv values varied from 2.40 to 2.54 (all clearly above the upper limit of 2.35 set forth in the claims), were processed and heat treated. Appellants’ Table II shows that all seven heats contained sigma phase.

TABLE I

CHEMISTRY-WEIGHT PERCENT

Heat No. C Cr Ni Co Pe Mo Ti Al B Ny

D1-379-1 0.01 15.3 Bal. 17.9 — 4.5 3.6 4.7 ’ 0.023 2.53

D1-379-2 0.04 15.3 Bal. 17.9 — 4.6 3.6 4.7 0.022 2.54

D1-380-1 0.06 15.3 Bal. 17.5 1.0 4.6 3.6 4.7 0.021 2.51

D1-380-2 0.06 15.1 Bal. 17.4 3.5 4.5 3.5 4.6 0.020 2.40

D1-382 0.06 15.3 Bal. 18.5 — 4,3 3.5 4.4 0.019 2.47

D1-383 0.06 15.2 Bal. 17.7 — 4.3 3.6 4.4 0.020 2.43

D1-386 0.06 15.3 Bal. 18.1 — 4.7 3.4 4.6 0.021 2.49

TABLE II

Approximate Heat No. w/o Sigma

D1-379-1 1.4

Dl-379-2 0.9

Dl-380 — 1 0.4

Dl-380-2 0.05

Dl-382 0.05

Dl-383 0.3

Dl-386 0.3

The affidavit states that “any amount of sigma phase is deleterious and undesirable because of the susceptibility to embrittlement failure following exposure to high temperature.”

The Board

The board agreed with the examiner that the claimed alloys were prima facie obvious from the prior art, noting that there was no substantial disagreement that both Pohl-man et al. and Lamb disclose alloys having compositional limits overlapping those of the claimed alloys. Although disagreeing with the examiner’s contention that there was no evidence to support the statement in the Boesch affidavit that “any amount of sigma phase is deleterious and undesirable,” it agreed with the examiner that the Boesch affidavit was insufficient to overcome the prima facie case of obviousness because there was no evidence showing:

(1) the precise amounts of sigma-like phase present in compositions containing Appellants’ claimed components balanced to provide Nv values just inside versus just outside Appellants’ claimed “about 2.35” Nv limits; and (2) direct comparisons of sufficient mechanical properties of those compositions within and without the claimed limit, to demonstrate the alleged critical correlation of Nv limit with sigma phase content.[⁸]

The board also said that the showing (in the specification, set forth infra) did not establish the asserted criticality in selection of the components of the alloys according to the claimed Nv formula, because the alloys failed to meet the claimed compositional and Nv value requirements.

OPINION

The Prima Facie Case

Each of the ranges of constituents in appellants’ claimed alloys overlaps ranges disclosed by Pohlman et al. and Lamb. Appellants, citing In re Waymouth, 499 F.2d 1273, 182 USPQ 290 (Cust. & Pat.App.1974), argue that neither of the cited prior art references recognizes the problem solved by them and, therefore, cannot render the claims obvious. Upon examination of the prior art references, we do not agree. Appellants admitted in their specification that:

It has been postulated according to Pauling’s theory that the criterion for the formation of sigma phase is based upon the number of electron vacancies (Nv) in the bonding orbitals of the elements involved. Based thereon, other investigators have derived an empirical equation which includes the elements chromium, molybdenum, manganese, iron, cobalt and nickel. It is to be noted, however, that the nickel base alloys to which reference is made in the present invention relate to an iron-free or low-iron composition, with only incidental amounts of an element such as manganese, and are hardened by the aluminum and titanium rich interme-tallic compound gamma prime.

U.S. patent No. 3,837,838 (’838), filed December 18, 1972, and issued September 24, 1974, was introduced into evidence by appellants and further illuminates what is meant by “Pauling’s theory”:

As described in an article by Linus Pauling entitled “The nature of intera-tomic forces in metals,” published in Physical Review, 54:899, 1938, in a given metallic atom, the outermost orbitals, termed the bonding orbitals, are occupied by the bonding electrons responsible for bonding the atom to its neighboring metallic atoms. At a given instant in time and on the average, the bonding orbitals are only partially occupied by the bonding electrons. Such partial occupation means that the outer orbitals are partially vacant of electrons or possess an “electron hole.” The total average number of vacant orbitals in a given metallic atom is called the electron hole number of the metal (Nv). The average electron hole number (Nv) is the resultant of adding all Nv for the participating elements in the alloy - matrix. The higher the Nv of a given Co-Cr-Ni alloy the higher the chance for the precipitation of embrit-tling phases. The quantities of metals consumed in precipitation do not enter in calculating Nv of the alloy matrix and hence do not participate in the formation of embrittling phases. A low Nv may thus be obtained by either choosing elements of low Nv to form an alloy or by using elements that will react in the alloy and precipitate out from the alloy matrix.

Accordingly, in carrying out this invention, I have selected an alloy-base for the system which possesses a low Nv, and have strengthened the alloy base by adding elements which will have minor or no effect on raising the Nv through controlling their percentage as solutes or by eliminating their effect on Nv by formation of compounds which precipitate out.

It appears from appellants’ specification that certain precipitate-hardened nickel base alloys, after being exposed to elevated temperatures for prolonged periods of time, suffered “from a marked and catastrophic decrease in room temperature ductility and a marked increase in the rate of creep deformation.” It was observed that other nickel base alloys having the same percentage ranges of components did not suffer such deleterious changes. The cause of the problem was believed to be the formation of an embrittling phase (“sigma”). As early as 1938, however, it was known that the higher the Nv value of a Co-Cr-Ni alloy, the higher the chance for precipitation of em-brittling phases; also, that the quantities of metals consumed in precipitation did not enter into calculating the Nv value of an alloy matrix. We are persuaded that one of ordinary skill in the art would have been guided by these principles.

In the above-quoted passage from ’838, we note that lowering the Nv value of a Co-Cr-Ni alloy and deletion of the metals not consumed in precipitation from the Nv calculation are expressly suggested. Considering, also, that the composition requirements of the claims and the cited references overlap, we agree with the Solicitor that the prior art would have suggested “the kind of experimentation necessary to achieve the claimed composition, including the proportional balancing described by appellants’ Nv equation.” This accords with the rule that discovery of an optimum value of a result effective variable in a known process is ordinarily within the skill of the art. In re Antonie, 559 F.2d 618, 195 USPQ 6 (Cust. & Pat.App.1977); In re Aller, 220 F.2d 454, 42 CCPA 824, 105 USPQ 233 (1955). Accordingly, we conclude that a prima facie case of obviousness has been established.

Unexpected Results

It is well settled that a prima facie case of obviousness may be rebutted “where the results of optimizing a variable, which was known to be result effective, [are] unexpectedly good.” In re Antonie, supra, 559 F.2d at 620, 195 USPQ at 8-9, and cases cited therein. It is also well settled that proof of unexpected properties may be in the form of direct or indirect comparative testing of the claimed compounds (here, alloys) and the closest prior art. In re Payne, 606 F.2d 303, 316, 203 USPQ 245, 256 (Cust. & Pat.App.1979), and cases cited therein.

A. Creep Tests

Table V, set forth in appellants’ specification and shown below, compares four examples of the claimed alloys with one example (6-3211) of a prior art alloy and is intended to show that the measured creep of the claimed alloys is unexpectedly less than that of the prior art.

TABLE V

Creep Tests at 1500 ° F and 37,000 psi

Alloy No. Sample Removed After (Hours) Measured Creep (inches per inch)

2-1422 1567.8 0.008

2-1423 1500.4 0.004

2-1425 1504.5 0.010

Alloy No. Sample Removed Measured Creep After (Hours) (inches per inch)

2-1426 1500.4 0.004

6-3211 1505.1 0.034

The measured creep of 6-3211 — an alloy, appellants note, having “chemistries” within those of the references — is in excess of three to eight times greater than the creep of the claimed alloys.

The composition and Nv values of the alloy heats in Table V are as follows:

Element, Weight %

Alloy No. A1 Ti Mo Cr Co B Ni Nv Value

2-1422 0.07 4.20 3.23 4.70 14.7 18.0 0.030 bal. 2.32

2-1423 0.06 4.37 3.45 4.45 14.6 17.6 0.028 bal. 2.36

2-1425 0.06 3.91 2.98 4.40 14.8 17.5 0.028 bal. 2.21

2-1426 0.05 4.20 3.19 4.50 14.5 17.5 0.030 bal. 2.27

6-3211 0.06 4.43 3.54 4.95 15.2 18.8 0.030 bal. 2.51

Although it is apparent that the molybdenum content of 6-3211 exceeds the maximum content of the claimed alloys by 0.15%; it is clearly within the ranges of the Pohl-man et al. and Lamb alloys.

However, we are not persuaded that the Table V data are commensurate in scope with appellants’ claims. In re Greenfield, 571 F.2d 1185, 1189, 197 USPQ 227, 230 (Cust. & Pat.App.1978).⁹ Appellants claim broad ranges of elements, but the weight percent of elements in the four examples of the claimed alloys vary by relatively minor amounts. For example, the entire claimed range of carbon is .18 percent, but the tested range is only .02 (.07 minus .05); the claimed cobalt range is 4.8, but the test range is only 1.3. There is no evidence showing whether other alloys encompassed by appellants’ broad claims and having elements varying by relatively major amounts also exhibit a low creep rate.

B. Ductility Test

Appellants’ Table VI, set forth in their specification, compares the room temperature ductility of one heat of the claimed alloy (2-1426) and one heat of an alloy (6-3266) which appellants state has “chem-istries” within those of the references.

TABLE VI

Room Temperature Tensile Tests

0.2% Alloy U.T.S. No._Condition_psi Offset Y.S. Elong. R.A. Nv (psi)_(%) (%) Value

2-1426 As-heat-treated 204,000 140.000 16.9 15.0 2.27

2-1426 As-heat-treated 157,000 + exposed 5000 hrs. at 1500 °F 100.000 16.1 14.1 2.27

6-3266 As-heat-treated 194,500 136,800 14.0 13.7 2.52

6-3266 As-heat-treated 150,500 + exposed 5000 hrs. at 1500 °F 117,500 5.0 5.5 2.52

The marked decrease in room temperature ductility (Elong.) after prolonged elevated temperature exposure of the prior art alloy (6-3266), compared to the claimed alloy’s (2-1426) essentially unchanged ductility, is contended to show an unexpected result, as was the improvement in measured creep discussed earlier. However, for the same reason that the measured creep test data of Table V are not persuasive of unexpected results, we do not regard the tensile test data of Table VI, comparing only one heat of a claimed alloy, sufficient to rebut the prima facie case of obviousness of the claimed invention.

C. Absence of Sigma Phase

Throughout prosecution appellants have maintained that the claims define “a nickel base alloy which can be manufactured in a consistent way to remain free from a tendency to form plate-like sigma phase.” The “essential concept of the present invention [is] to maintain the average number of electron vacancies at a value not exceeding about 2.35.” Whereas the Pauling theory teaches that a low Nv value means reduced chances for sigma phase, appellants allege that alloys meeting their composition and Nv value requirements are free from sigma phase.

As related earlier, the Boesch affidavit shows that sigma phase is present in seven alloy examples, all of which meet the composition requirements but exceed the Nv value requirement of the claimed alloys. However, this affidavit contains no examples of claimed alloys showing the absence, or presence, of sigma. The remainder of the record reveals only a single example of the claimed alloy, which shows the absence of sigma.¹⁰ Appellants’ specification includes a photomicrograph of Table V alloy heat 2-1422, which clearly shows the absence of sigma; also, a photomicrograph of Table V alloy heat 6-3211, which shows the presence of sigma. We note again that the prior art teaches that reduction of the Nv value reduces the chances of sigma phase in the alloy. Here appellants tested only one example of a low Nv value alloy and found no sigma — a result consistent with both the prior art teaching and appellants’ allegation that their claimed alloys are “totally free from sigma phase.”¹¹ Under such circumstances, test results involving a single alloy within the broad range claimed are not sufficient to support appellants’ allegation of what would, from the prior art, be unexpected.¹²

In view of the foregoing we hold that appellants have failed to rebut the prima facie case of obviousness.

The decision of the board is affirmed.

AFFIRMED.

1 . Serial No. 587,776 was filed on June 17, 1975.

2 . U.S. patent No. 3,147,155, issued September 1, 1964.

3 . U.S. patent No. 3,457,066, issued July 22, 1969.

4 . Nv refers to the average electron vacancy concentration per atom in the matrix of the alloy.

5 . Appellants state that the overall variation in Nv due to chemical uncertainty is ± 0.025 so that in reality the Nv value of about 2.35 may actually extend from 2.32 to 2.38.

6 . Appellants’ specification states that A% “refers to the atomic percent of the element so described.”

7 . Creep is the permanent deformation of a metal that occurs as a result of prolonged compression or extension at or near room temperature. The Condensed Chemical Dictionary 228 (8th ed. 1971).

8 . The board agreed with the examiner that “there [was no evidence showing] that an alloy having an Nv number of 2.35 is free of any amount of sigma phase, or what the sigma phase content and properties are of an alloy having an Nv number of 2.36 which is close to but outside the Nv requirement.”

9 . It is unnecessary to decide whether 6-3211 is the “best prior art.” See In re Malagari, 499 F.2d 1297, 1302-03, 182 USPQ 549, 552-53 (Cust. & Pat.App.1974).

10 . Thus, appellants have again failed to show test data commensurate in scope with the broad claims.

11 . We agree with the board that the six United States patents ((1) No. 4,093,474, issued June 6, 1978; (2) No. 4,083,734, issued April 11, 1978; (3) No. 3,930,904, issued January 6, 1976; (4) No. 3,837,838, issued September 24, 1974; (5) No. 3,816,110, issued June 11, 1974; and (6) No. 3,767,385, issued October 23, 1973) introduced into the record by appellants “do support the assertion in the Boesch affidavit that ‘any amount of sigma phase’ is undesirable.” Therefore, we have limited our analysis to the issue of the existence of sigma phase and have not extended it to include the effect of varying amounts of sigma phase.

12 . Where it is alleged that a certain technique for flipping coins would always produce “heads,” one would hardly be persuaded by a single toss of a coin which resulted in a showing of “heads.”