Opinion for the court filed by Circuit Judge LOURIE.
Opinion dissenting in part filed by Circuit Judge BRYSON.
LOURIE, Circuit Judge.
This patent appeal arises as one of a dying breed of interference proceedings before the United States Patent and Trademark Office Patent Trial and Appeal Board (“the Board”) between Thomas Beckmann and Alexander Massner (collectively, “Beckmann”), the named inventors of U.S. Patent 7,584,605 (“the '605 patent”), and Harén S. Gandhi, John Vito Cavataio, Robert Henry Hammerle, and Yisun Cheng (collectively, “Gandhi”), the applicants of U.S. Patent Application 12/877,901 (“the '901 application”). Beck-mann appeals from the Board’s decision, which held that all of Beckmann’s claims involved in the interference, viz., claims 1-4 of the '605 patent, are unpatentable over the cited prior art, and that claims 5, 7, 8, and 10-17 of Gandhi’s '901 application satisfy the written description requirement. See Beckmann v. Gandhi Interference No. 105, 822, Paper No. 114 (P.T.A.B. July 8, 2013) (“Decision on Motions”)-, Beckmann v. Gandhi Interference No. 105,822, Paper No. 117 (P.T.A.B. Mar. 9, 2015). (“Rehearing Decision ”).
For the reasons that follow, we affirm the Board’s determination that claims 3 and 4 of the '605 patent are unpatentable as obvious over the cited prior art, and that the '901 application contains an adequate written description of the “transitioning” limitation of claims 12-17, as well as the “minimizing the oxygen content” limitation of claims 7,10, and 16. However, because the Board erred in construing claim 1 of the '605 patent as not requiring a separate and distinct third “supplying” step, we vacate its determination that [952]*952claims 1 and 2 of the '605 patent are unpatentable over the cited prior art and remand for further proceedings under the proper claim construction. Moreover, because the Board’s finding that the '901 application contains an adequate written description of the “minimizing the oxygen content ... prior to a rich cycle” limitation of claims 5, 8,11, and 17 is unsupported by substantial evidence, we vacate that finding and remand for further proceedings consistent with this opinion.
Background
I
The technology at issue relates to methods of reducing nitrogen oxide (“NOx”) emissions from internal combustion engines. When an engine operates in a lean mode, a reduced amount of fuel is provided to the cylinders, which improves fuel efficiency. But operating in the lean mode produces an exhaust gas with an excess of oxidizing constituents (“lean exhaust gas”), which decreases the effectiveness of the exhaust gas purification system, leading to NOx emissions. In contrast, a “rich exhaust gas” is one that has an excess of reducing constituents.
According to Beckmann, a prior-art solution for reducing NOx emissions involved operating the engine alternately in lean and rich modes to produce lean and rich exhaust gases. Appellants’ Br. 10. In the engine, an NOx catalytic converter (or “lean NOx trap”) stores NOx produced during the lean cycle, and then converts the stored NOx to ammonia (“NH3”) during the rich cycle. A downstream selective catalytic reduction (“SCR”) catalytic converter then stores the NH3 and reacts it with any NOx released from the NOx catalytic converter to produce nitrogen gas N2 for release into the atmosphere.
[[Image here]]
J.A. 1453.
When the engine switches from lean to rich mode, however, the sudden change in exhaust gas composition from lean to rich causes the direct mixing of oxidizing and reducing constituents inside the cavities of the NOx catalytic converter. '605 patent col. 2 11. 15-27. Such mixing impedes the efficient reduction of NOx to NH3, and also causes the release of a large amount of NOx from the NOx catalytic converter. Id. col. 2 11. 27-35. That problem was known as “NOx spike.”
Beckmann’s '605 patent purportedly solves the NOx spike problem by interposing a third operating mode between the lean and rich modes. Id. col. 2 11. 35-45. The third operating mode takes place after the lean mode and before the rich mode, id. col. 2 11. 7-10, during which the NOx catalytic converter is supplied with an exhaust gas that has a lower content of oxidizing constituents than the lean exhaust gas and a lower content of reducing constituents than the rich exhaust gas (“intermediate exhaust gas”), id. col. 2 11. .10-14. In the third operating mode, the intermediate exhaust gas replaces the lean exhaust gas inside the cavities of the NOx catalytic converter, which mitigates the NOx spike problem. Id. col. 2 11. 46-61.
The '605 patent has four claims, among which claims 1 and 3 are independent. [953]*953Claims 2 and 4 depend from claims 1 and 3, respectively. Claim 1 reads as follows:
1. A method for purifying the exhaust gas from an internal combustion engine having an exhaust-gas purification system including a nitrogen oxide storage catalytic converter and an SCR catalytic converter downstream of the nitrogen oxide storage catalytic converter, comprising the steps of:
supplying the nitrogen oxide storage catalytic converter with exhaust gas containing an excess of oxidizing constituents [“the lean step”];
supplying the nitrogen oxide storage catalytic converter with exhaust gas containing an excess of reducing constituents [“the rich step”]; and
supplying the nitrogen oxide storage catalytic converter, between the oxidizing constituents supplying step and the reducing constituents supplying step, ioith an exhaust gas which has a lower content of oxidizing constituents than in the oxidizing constituents supplying step and a lower content of reducing constituents than in the reducing constituents supplying step [“the intermediate step”],
wherein the step between the oxidizing constituents supplying step and the reducing constituents supplying step is terminated at the earliest when the nitrogen oxide storage catalytic converter is predominantly filled by exhaust gas delivered in step between the oxidizing constituents supplying step and the reducing constituents supplying step.
Id. col. 1311.15-39 (emphases added).
Claim 3 recites the same preamble and lean and rich steps as claim i, but a different intermediate step and wherein clause, which are reproduced below.
3. ...
supplying the nitrogen oxide storage catalytic converter, between the oxidizing constituents supplying step and the reducing constituents supplying step, for a predetermined period with a constant exhaust gas composition which has a lower content of oxidizing constituents than in the oxidizing constituents supplying step and a lower content of reducing constituents than in the reducing constituents supplying step [“the intermediate step”],
wherein in step of supplying the nitrogen oxide storage catalytic converter between the oxidizing constituents supplying step and the reducing constituents supplying step, an air/fuel ratio which set to control exhaust gas composition is set to be slightly greater than 1, such that the oxidizing constituents in the exhaust gas have an oxygen excess of 1% or less.
Id. col. 1411.11-36 (emphases added).
II
In 2010, Gandhi filed the '901 application, which is a continuation in a series of applications, including U.S. Patent Application 10/065,470 (“the '470 application”). The '470 application was issued as U.S. Patent 7,332,135 (“Gandhi '135”), which is prior art to Beckmann’s '605 patent. Gandhi’s '901 application and Gandhi '135 share the same specification in relevant part.
In the Background section, the '901 application describes the NOx spike problem that “occurs during the short. period in which the NOx trap transitions from lean to rich,” and states that “FIG. la [shows] that during this lean-rich transition, NOx spikes, the large peaks of unreacted NOx,” account for significant NOx emissions. [954]*954J.A. 502. Then, in the Summary of the Invention section, the '901 application states that the NOx spike problem was solved “with the use of an NH3-SCR catalyst placed downstream of the lean NOx adsorber catalyst.”1 J.A. 503. The '901 application explains that “[t]he advantage of the catalyst system of this invention is the use of a combination of a lean NOx trap and an NH3-SCR catalyst.” J.A. 505. Additionally, in paragraph 41, the '901 application states that: “For this invention, the lean NOx trap is optimized for ammonia generation by removing oxygen storage capacity (OSC) and thereby enhancing the rich cycle and thus creating a greater quantity of ammonia for reaction with the NOx in the downstream NH3-SCR catalyst.” Id. (emphasis added).
When filing the '901 application, Gandhi added claims 3-17, seeking to provoke an interference with Beckmann. Gandhi’s claim 3 was copied from, and thus is identical to, claim 1 of the '605 patent. Although none of Gandhi’s other claims, viz., claims 4-17, is completely identical to any of Beckmann’s claims, Gandhi stated that claims 4-11 “are substantially identical to, and have been copied from,” Beckmann’s claim 1, and that claims 12-17 “correspond substantially to” Beckmann’s claim 1. J.A. 515.
Claims 5, 8, 11, and 17 of the '901 application each contain a “minimizing the oxygen content ... prior to a rich cycle” limitation. J.A. 507-08, 510-13. Claim 5 is illustrative and reads as follows:
5, A method for reducing pollutants in the exhaust gas of an engine having a system including a nitrogen oxide ad-sorber and a NHs-SCR catalyst downstream of the nitrogen oxide adsorber, comprising the steps of:
supplying the nitrogen oxide adsorber with exhaust gas while the engine operates under lean conditions, wherein the exhaust gas is a lean exhaust gas;
transitioning the engine operations from lean to rich conditions[;]
supplying the nitrogen oxide adsorber with exhaust gas during said transitioning;
minimizing the oxygen content of the nitrogen oxide adsorber prior to rich conditions to facilitate the reduction of NOx to NH3;
completing the transition from lean to rich conditions; and
supplying the nitrogen oxide adsorber with exhaust gas while the engine operates under rich conditions, wherein the exhaust gas is a rich exhaust gas.
J.A. 507, 510 (emphasis added). Claims 7, 10, and 16 each contain the limitation “the oxygen content of the lean NOx trap is minimized to facilitate the reduction of NOx to NH3,” but those claims do not include any “prior to a rich cycle” language. J.A. 508, 511-12. Claims 8, 11, and 17 depend from claims 7, 10, and 16, respectively.
Claim 12 and its dependent claims 13-17 require “transitioning” between a lean exhaust gas supplying step and a rich exhaust gas supplying step and “completing said transition.” Claim 12 reads as follows:
12. A method for purifying the exhaust gas from an engine, comprising:
supplying a zoned catalyst system with a lean exhaust gas, wherein said zoned catalyst system comprises a first zone of lean NOx trap and a second zone of SCR catalyst downstream of the first zone;
[955]*955supplying the zoned catalyst system with a rich exhaust gas;
transitioning between said lean exhaust gas supplying step and said rich exhaust gas supplying step, wherein exhaust gas is supplied to said zoned catalyst system during said transition; and
completing said transition from said lean exhaust gas supplying step to said rich exhaust gas supplying step.
J.A. 508, 512 (emphases added).
Ill
In July 2011, the Board declared Interference No. 105,822 between Gandhi’s '901 application and Beckmann’s '605 patent. The interference involves a sole count corresponding to claims 1-4 of the '605 patent and claims 3-17 of the '901 application. The Count is identical to Beckmann’s claim 1 and Gandhi’s claim 3. When declaring the interference, the Board accorded Gandhi the benefit of the '470 application, filed in October 2002, and accorded Beckmann the benefit of an application filed in October 2003.2 Id. Based on those filing dates, the Board designated Gandhi as the senior party.
The parties then filed several preliminary motions, including: (1) Beckmann’s motion alleging that Gandhi’s claims 3-5, 7, 8, and 10-17 are unpatentable for lack of written description support (“Beckmann Motion 1”); (2) Beckmann’s motion alleging that Gandhi’s claims 3, 4, 6, 9, and 12 are unpatentable in view of European Patent Publication EP 0 814 241 (“Kinugasa '241”), U.S. Patent 6,109,024 (“Kinugasa '024”), and U.S. Patent Publication 2003/0056499 (“Binder '499”) (“Beckmann Motion 2”); and (3) Gandhi’s motion alleging that Beckmann’s claims 1-4 are unpat-entable in view of Gandhi '135 and U.S. Patent 6,604,504 (“Surnilla '504”) (“Gandhi Motion 1”). The Board decided the parties’ motions in the following order: (1) Beckmann Motion 1; (2) Beckmann Motion 2; and (3) Gandhi Motion 1. Decision on Motions at 4.
First, the Board denied Beckmann Motion 1, finding that Gandhi’s claims 3-5, 7, 8, and 10-17 are not unpatentable for lack of written description support. Id. at 7-24. Beckmann raised three separate written description challenges: (1) that Gandhi’s claims 3 and 4, copied from Beck-mann, must be construed in light of the '605 patent as requiring a third exhaust gas supplying step that is (a) separate and distinct from the lean and rich exhaust gas supplying steps and (b) controlled in terms of exhaust gas composition and duration, and that Gandhi’s '901 application lacks an adequate written description of such a third supplying step; (2) that Gandhi’s '901 application lacks an adequate written description of the “transitioning” limitation of claims 12-17; and (3) that Gandhi’s '901 application lacks an adequate written description of the “minimizing the oxygen content” limitation of claims 5, 7, 8, 10, 11, 16, and 17.
With respect to Gandhi’s claims 3 and 4, the Board recognized the differences between Beckmann’s and Gandhi’s disclosures, and agreed with Beckmann that Gandhi’s '901 application does not contain a written description of a separate and distinct third exhaust gas supplying step that is controlled in terms of exhaust gas composition and duration. Id. at 14-15. But the Board concluded that Gandhi’s [956]*956claims 3 and 4, as construed in light of Beckmann’s '605 patent, see Agilent Techs., Inc. v. Affymetrix, Inc., 567 F.3d 1366, 1375 (Fed.Cir.2009), do not require such a third supplying step. Decision on Motions at 15-16.
Instead, the Board concluded that the claims only require that, between the lean and rich exhaust gas supplying steps, the NOx catalytic converter “is supplied with an exhaust gas having less oxidizing constituents than lean operations and less reducing constituents than rich conditions, and that this supplying step terminates at the earliest when the lean NOx trap is predominantly full of exhaust gas supplied during this step.” Id. at 16. The Board also noted that the differences between Beckmann’s claim 1 (which is identical to Gandhi’s claim 3) and Beckmann’s claim 3 (which requires “a constant exhaust gas composition” and “a predetermined period” for the third supplying step) further support its construction of Gandhi’s claim 3. Id. at 16-19.
The Board then found that the '901 application provides an adequate written description of Gandhi’s claims 3 and 4 because it describes “the short period in which the NOx trap transitions from lean to rich,” and because the claims only require “the transition between lean and rich supplying steps in which an exhaust gas having less oxidizing constituents than lean operations and less reducing constituents than rich conditions is supplied, and that this supplying step (transition step) terminates at the earliest when the lean NOx trap is predominantly full of the exhaust gas supplied.” Id. at 19-20.
With respect to Gandhi’s claims 12-17, the Board reasoned that claim 12 does not require “a separate and distinct transition step between the supply of lean and rich exhaust gases,” and that the claim “merely requires transitioning between lean and rich conditions and completing the transition to rich conditions, as described by Gandhi’s disclosure.” Id. at 21. The Board again relied on the '901 application’s description of “the short period in which the NOx trap transitions from lean to rich” and its Figure la to find that the '901 application adequately describes “transitioning” between lean and rich conditions and “completing” that transition. Id. at 21-22.
With respect to Gandhi’s claims 5, 7, 8, 10, 11, 16, and 17, the Board noted that those claims require “the minimization [of] the oxygen content in the NOx catalytic converter prior to the rich cycle.” Id. at 23 (emphasis in original). The Board found that the '901 application, especially its paragraph 41, which describes “removing oxygen storage capacity” of the lean NOx trap, adequately describes the “minimizing” limitation. Id.
Next, the Board granted Beckmann Motion 2 in part, holding that Gandhi’s claims 3, 4, 6, and 9, as construed in light of Gandhi’s '901 application, see Agilent, 567 F.3d at 1375, are anticipated by, or would have been obvious over, Kinugasa '241 and Kinugasa '024. Decision on Motions at 24-35. The Board denied the motion as to Gandhi’s claim 12, however, finding that the cited reference, Binder '499, is not prior art to Gandhi’s claim 12. Id. at 35-41.
Pursuant to 37 C.F.R. § 41.207(c), the Board then considered whether Beckmann rebutted the presumption of cross-applicability of the prior art to its own claims, viz., claims 1-4 of the '605 patent. The Board concluded that Beckmann failed to do so as to claims 1 and 2. Id. at 42-44. Specifically, the Board noted that Beck-mann sought to overcome the presumption by making the same claim construction argument, namely, that Beckmann’s claims, as construed in light of the '605 [957]*957patent, require a separate and distinct third exhaust gas supplying step that is controlled in terms of exhaust gas composition and duration. Recognizing that Beckmann’s claims 3 and 4 indeed require a constant exhaust gas composition and a predetermined period for the third exhaust gas supplying step, the Board concluded that Beckmann successfully overcame the presumption as to those claims. But because the Board, when deciding Beckmann Motion 1, rejected Beckmann’s proposed construction of Gandhi claim 3 and Beck-mann claim 1, the Board concluded here that Beckmann failed to rebut the § 41.207(c) presumption as to its claims 1 and 2. The Board therefore held that claims 1 and 2 of the '605 patent are unpatentable over Kinugasa '241, Kinuga-sa '024, and Binder '499.
Last, the Board granted Gandhi Motion 1 as to Beckmann’s claims 3 and 4, concluding that those claims would have been obvious over Gandhi '135 and Sumilla '504. Id. at 45-52. The Board acknowledged that claims 3 and 4 require a constant exhaust gas composition, a predetermined period, and a specific air/fuel ratio for the third exhaust gas supplying step. Id. at 48. As discussed below, the Board implicitly found that those claims require a separate and distinct third step. But it found that those features were taught or suggested by Sumilla '504. Id¡. at 48-50. In particular, the Board found that Surnil-la '504 taught a stepwise transition from lean to rich conditions, which resulted in constant intermediate exhaust gas compositions over a predetermined period and reduced NOx emissions. Id. at 49-50. The Board reasoned that one of ordinary skill would have been motivated to implement the stepwise transition taught by Sumilla '504 in the system disclosed by Gandhi '135 to further reduce NOx emissions. The Board therefore concluded that claims 3 and 4 of the '605 patent would have been obvious over Gandhi '135 and Surnilla '504.
Gandhi also argued that Beckmann’s claims 1 and 2 are unpatentable over Gandhi '135 and Surnilla '504. But the Board did not reach that issue because it found Beckmann’s claims 1 and 2 unpatentable over Kinugasa '241, Kinugasa '024, and Binder '499. Id. at 44.
Because the Board found all of Beck-mann’s involved claims to be unpatentable, it terminated the interference and entered judgment. Beckmann v. Gandhi Interference No. 105,822, 2013 WL 3788515 (P.T.A.B. July 8, 2013). Beckmann timely requested rehearing, which the Board denied. Rehearing Decision at 13. Beck-mann then appealed to this court. We have jurisdiction under 28 U.S.C. § 1295(a)(4)(A).
Discussion
We review the Board’s legal determinations de novo, In re Elsner, 381 F.3d 1125, 1127 (Fed.Cir.2004), and its factual findings for substantial evidence, In re Gartside, 203 F.3d 1305, 1316 (Fed.Cir.2000). A finding is supported by substantial evidence if a reasonable mind might accept the evidence to support the finding. Consol. Edison Co. v. NLRB, 305 U.S. 197, 229, 59 S.Ct. 206, 83 L.Ed. 126 (1938).
On appeal, Beckmann argues that the Board erred in construing claims 1-4 of the '605 patent and in determining that those claims are unpatentable over the cited prior art. Beckmann also argues that the Board erred in construing claims 12-17 of Gandhi’s '901 application and in finding that the '901 application provides adequate written description support for the “transitioning” limitation of those claims. Finally, Beckmann argues that the Board erred in finding that the '901 application contains an adequate written [958]*958description of the “minimizing the oxygen content” limitation of claims 5, 7, 8, 10, 11, 16, and 17.3 We address each of those arguments in turn.
We first consider whether the Board erred in construing claims 1 and 2 of the '605 patent. We review the Board's ultimate claim construction de novo and its underlying factual determinations involving extrinsic evidence for substantial evidence. Microsoft Corp. v. Proxyconn, Inc., 789 F.3d 1292, 1297 (Fed.Cir.2015). Here, because the intrinsic record fully determines the proper construction, we review the Board’s construction de novo. See id.
“Interference counts are given the broadest reasonable interpretation possible.” Davis v. Loesch, 998 F.2d 963, 968 (Fed.Cir.1993); see also Yorkey v. Diab, 605 F.3d 1297, 1300-01 (Fed.Cir.2010) (stating that claims are given their broadest reasonable interpretation in an interference proceeding). “In determining the true meaning of the language of the count, the grammatical structure and syntax thereof may be instructive.” Credle v. Bond, 25 F.3d 1566, 1571 (Fed.Cir.1994).
When deciding Beckmann Motion 1, the Board construed Gandhi’s claim 3, which is identical to Beckmann’s claim 1, in light of the '605 patent specification, see Agilent, 567 F.3d at 1375, and rejected Beckmann’s argument that the claim requires a separate and distinct third exhaust gas supplying step that is controlled in terms of exhaust gas composition and duration. Decision on Motions at 15-16. Based on that construction, when deciding Beckmann Motion 2, the Board found Beckmann’s claim 1 and dependent claim 2 to be unpatentable over the cited prior art. Id. at 42-44.
Beckmann argues that the Board erred in construing its claim 1 as not requiring a separate and distinct third exhaust gas supplying step. According to Beckmann, the structure and wording of the claim demonstrate that each of the three “supplying” steps is distinct from the other two. Beckmann contends that the Board’s construction improperly reads the third supplying step out of the claim and is inconsistent with the '605 patent’s specification, which describes three distinct operating modes (viz., lean, intermediate, and rich), solves the NOx spike problem by supplying an intermediate exhaust gas between the lean and rich modes, and distinguishes the claimed method from prior-art methods that involved only lean and rich modes and direct transitions between those two modes.
Beckmann additionally argues that the Board erred in construing claim 1 as not requiring the third exhaust gas supplying step to be controlled in terms of both exhaust gas composition and duration. According to Beckmann, the claim specifies the exhaust gas composition, viz., an intermediate exhaust gas with a lower level of oxidizing constituents than the lean exhaust gas and a lower level of reducing constituents than the rich exhaust gas, as well as the duration of the third supplying step, viz., that the third step be “terminat[959]*959ed” when the NOx catalytic converter is “predominantly filled” with the intermediate exhaust gas. Beckmann contends that those limitations imply control over the exhaust gas composition and duration. Finally, Beckmann argues that the Board misapplied the principles of claim differentiation when analysing the differences between Beckmann claim 1 and Beckmann claim 3.
Gandhi responds that the Board’s construction is consistent with the plain meaning of the claim language, as well as the specification of the '605 patent. According to Gandhi, the Board properly declined to import limitations from the specification into the claim. Gandhi contends that Beckmann chose to draft its claim broadly, and that nothing in the grammar, structure, or wording of the claim requires that there be three separate and distinct exhaust gas supplying steps or that the exhaust gas composition be controlled over time during the third step. Gandhi also responds that the Board’s construction does not read the third step out of the claim because the construction clearly requires that the third supplying step occur between the first and second supplying steps. Finally, Gandhi responds that the Board did not misapply claim differentiation by recognizing that Beckmann claim 1 is broader than Beckmann claim 3.
We agree with Beckmann that its claim 1 requires a third exhaust gas supplying step that is separate and distinct from the lean and rich exhaust gas supplying steps, but we conclude that the claim does not require that the third supplying step be controlled in terms of exhaust gas composition and duration.
Beginning with the language of the claim, claim 1 is directed to a method “comprising the steps of....” '605 patent col, 13 1. 19 (emphasis added).- The claim then lists three parallel “supplying” clauses, separated by line indentation, id. col. 13 11. 20-32, and followed by a “wherein” clause that refers back to the three “supplying” clauses as steps, id. col. 13 11. 33-39 (“wherein the step between the oxidizing constituents supplying step and the reducing constituents supplying step ... delivered in step between the oxidizing constituents supplying step and the reducing constituents supplying step ” (emphases added)). The structure of the claim thus indicates that the claimed method comprises three different supplying steps.
Moreover, both the third “supplying” clause and the “wherein” clause require the third supplying step to be between the first step (the oxidizing constituents supplying step or the lean step) and the second step (the reducing constituents supplying step or the rich step). Id. col. 13 11. 26-39. Accordingly, the claim imposes a sequential requirement that the third step occur between the first and second steps.
The claim language also requires that the supplying steps be separate and distinct from each other. In each step, a different exhaust gas is supplied to the NOx catalytic converter: a lean exhaust gas is supplied in the first step; a rich exhaust gas is supplied in the second step; and an intermediate exhaust gas is supplied in the third step. Each step is therefore distinctly characterized by a different exhaust gas composition that is being supplied to the NOx catalytic converter. Moreover, the “wherein” clause requires that the intermediate exhaust gas supplying step, occurring between the lean and rich steps, be “terminated at the earliest when” the NOx catalytic converter is “predominantly filled by exhaust gas delivered” in the step between the lean and rich steps. Thus, the “wherein” clause imposes a further limitation that the intermediate supplying step does not end, and the next supplying step does not begin, until the [960]*960NOx catalytic converter is “predominantly filled” with the intermediate exhaust gas. Id. col. 13 11. 33-89.
Notably, the claim uses different terms, “supply! ]” and “fill[ ],” and those terms carry different meanings when read in the context of the surrounding claim language. For example, during the third supplying step, an intermediate exhaust gas is “supplied]” to the NOx catalytic converter, meaning that the intermediate exhaust gas is introduced through an inlet into the NOx catalytic converter. Inside the cavities of the NOx catalytic converter, that exhaust gas is then mixed with the existing exhaust gas, and shortly after the intermediate exhaust gas “predominantly fíll[s]” the NOx catalytic converter, the third supplying step terminates. This interpretation of “supply[ ]” and “fill! ]” is consistent with the plain meaning of the claim and the specification of the '605 patent. See, e.g., '605 patent col. 211.15-61, col. 4 1. 48-col. 5 1. 16, col. 7 1. 63-col. 8 1. 6, col. 8 11. 47-59.
In rejecting Beckmann’s proposed claim construction, the Board reasoned that “when operations are transitioned from lean to rich, there will be mixing between the leading edge of the rich exhaust gas and the lean exhaust gas preceding it such that intermediate exhaust gas will have a composition varying between lean and rich.” Rehearing Decision at 4. But the Board failed to properly consider the “wherein” clause, which requires that the intermediate exhaust gas supplying step not end, and the rich exhaust gas supplying step not begin, until the NOx catalytic converter is “predominantly filled” with the intermediate exhaust gas.
It is true that when an engine transitions directly from lean to rich, a natural consequence of that transition is the mixing of lean and rich exhaust gases at the interface to form a small amount of intermediate exhaust gas, which is then supplied transitorily into the NOx catalytic converter. But there is no evidence that such a small amount of intermediate exhaust gas would inevitably “predominantly fill[]” the NOx catalytic converter before the rich exhaust gas is beginning to be supplied into the converter. In those circumstances, the cavities of the NOx catalytic converter may still be predominantly filled with the lean, rather than the intermediate, exhaust gas. As the '605 patent specification explains, a direct transition from lean to rich modes leads to the direct mixing of lean and rich exhaust gases inside the cavities of the NOx catalytic converter, which causes the undesirable NOx spike problem. '605 patent col. 2 11. 21-27 (“Since the nitrogen oxide storage catalytic converter used is either a honeycomb body with passages passing through it or a bulk bed of shaped bodies, in the event of a sudden change in the exhaust-gas composition from oxidizing to reducing, the exhaust gases of different compositions become mixed with one another in the cavities formed by these catalyst converter structures.”).
We therefore conclude that the Board failed to properly consider the structure and wording of the claim and erred in construing claim 1 in an unreasonably broad manner as not requiring a separate and distinct third exhaust gas supplying step.
We agree with the Board, however, that claim 1 of the '605 patent does not require control over exhaust gas composition and duration of the third supplying step. Although the specification of the '605 patent describes various means for controlling the exhaust gas composition and duration of the exhaust gas supplying steps, see, e.g., '605 patent col. 2 1. 62-col. 3 1. 3, col. 10 1. 59-col. 11 1. 50, those descriptions are not part of claim 1. Indeed, the language of claim 1 does not specify any means by [961]*961which to control the exhaust gas composition and duration beyond the requirement that the NOx catalytic converter be “supplied]” with a certain type of exhaust gas in each of the supplying steps.
For the foregoing reasons, we conclude that Beckmann’s claims 1 and 2 require a third exhaust gas supplying step that is separate and distinct from the other two exhaust gas supplying steps, but that the claims do not require control over exhaust gas composition and duration of the third supplying step. Accordingly, we reverse the Board’s construction of claims 1 and 2 of the '605 patent.
We next consider whether the Board erred in construing claims 3 and 4 of the '605 patent. On appeal, Beckmann alleges that the Board erred in construing claim 3 and its dependent claim 4 as not requiring a separate and distinct third exhaust gas supplying step. Appellants’ Br. 20-28 (same argument as that raised for claims 1 and 2). Beckmann does not allege, however, that the Board erred by failing to construe claims 3 and 4 as requiring a third exhaust gas supplying step controlled in terms of both exhaust gas composition and duration. See id. at 28-30.
The Board did not formally construe claim 3 of the '605 patent, but the record shows that it recognized that claim 3, unlike claim 1, requires a third exhaust gas supplying step that has a constant exhaust gas composition, a specific air/fuel ratio, and a predetermined duration. Decision on Motions at 17, 19, 45, 48. Indeed, relying on those limitations of claim 3 not present in claim 1, the Board concluded that Beckmann overcame the presumption under 37 C.F.R. § 41.207(c) for claims 3 and 4, but not for claims 1 and 2. Id. at 43-44.
By correctly recognizing that claim 3 requires a third exhaust gas supplying step with a constant exhaust gas composition, a specific air/fuel ratio, and a predetermined duration, the Board implicitly construed claim 3 as requiring a third intermediate exhaust gas supplying step that is separate and distinct from the lean and rich steps. Beckmann’s argument to the contrary is without merit.
We therefore conclude that the Board did not err in construing claims 3 and 4 of the '605 patent.
We next consider whether the Board erred in concluding that claims 3 and 4 of the '605 patent would have been obvious over Gandhi '135 and Surnilla '504. Obviousness is a question of law based on underlying factual findings. In re Baxter, 678 F.3d 1357, 1361 (Fed.Cir.2012). Those factual inquiries include what a reference teaches, In re Beattie, 974 F.2d 1309, 1311 (Fed.Cir.1992), and the existence of a reason to combine references, In re Hyon, 679 F.3d 1363, 1365-66 (Fed.Cir.2012).
The Board found that Gandhi '135 discloses the lean and rich exhaust gas supplying steps, but does not disclose a third exhaust gas supplying step with a constant exhaust gas composition, a predetermined duration, and a specific air/fuel ratio as required by claim 3. Decision on Motions at 48. The Board nevertheless found that Surnilla '504 discloses or suggests those features. Id. The Board also found that one of ordinary skill in the art would have been motivated to combine those references to achieve the claimed invention. Id. at 51.
On appeal, Beckmann argues that Sur-nilla '504 does not disclose the exact composition of exhaust gas supplied to the lean NOx trap. According to Beckmann, Sur-[962]*962nilla '504 only discloses the composition of exhaust gas supplied to an “emission control device” 32 upstream to the'NOx trap 34. Beckmann argues that emission control device 32 is an unknown device, and that the Board erred in finding that component 32 is necessarily a three-way catalyst. Gandhi responds that substantial evidence supports the Board’s finding on what the prior art teaches and the Board’s finding that there would have been a reason to combine Gandhi '135 and Surnilla '504.
We conclude that substantial evidence supports the Board’s finding that Surnilla '504 teaches or suggests a third exhaust gas supplying step with a constant exhaust gas composition, a predetermined period, and a specific air/fuel ratio as required by claim 3. In particular, the Board correctly found that Surnilla '504 discloses controlling engine operations to achieve constant compositions of intermediate exhaust gas when transitioning between lean and rich conditions. Surnilla '504 col. 6 11. 8-37; id. fig.3 & fig.4. Surnilla '504 achieves this by sequentially stepping down the air/fuel ratio provided to each of four cylinders from a lean ratio to a rich ratio. Id. col. 4 1. 64-col. 5 1. 6; id. fig.2. As the Board correctly found, in the Surnilla '504 process, when three of the four cylinders are stepped down, the combined air/fuel ratio is slightly greater than 1, and at that point, the exhaust gas necessarily has an excess oxygen content of 1% or less. Further, the Board properly found that Surnilla '504 discloses that each of the down-steps involves a specific, predetermined wait period. Those factual findings are all supported by substantial evidence.
We are also unpersuaded by Beck-mann’s argument that the Board made a reversible error by finding that Surnilla '504 discloses that component 32 is a three-way catalyst. As the Board noted, Surnilla '504 explains that three-way catalysts are also referred to as emission control devices, id. col. 1 11. 31-32, and then refers to component 32 as an emission control device, id. col. 4 1. 52. Substantial evidence therefore supports the Board’s finding that emission control device 32 is a three-way catalyst.
■ It is true that Surnilla '504 also refers to component 34 as an emission control device, id., and later refers to 34 as an NOx trap, id. col. 6 11.11-12, suggesting that an emission control device could be an NOx trap. But for obviousness, we consider not only what a reference necessarily discloses, but also what that reference would have suggested to one of ordinary skill in the art. Here, in view of the teachings of Surnilla '504 as a whole, we conclude that the Board did not err in finding that Sur-nilla '504 teaches or suggests the limitations of claim 3 not otherwise disclosed by Gandhi '135.
Accordingly, for the foregoing reasons, the Board did not err in concluding -that claims 3 and 4 would have been obvious over Gandhi '135 and Surnilla '504.
IV
We next consider whether the Board erred in finding claims 1 and 2 of the '605 patent to be unpatentable over Kinugasa '241, Kinugasa '024, and Binder '499. When granting Beckmann Motion 2 in part, the Board determined that Beck-mann’s claims 1 and 2 are also unpatentable because Beckmann failed to rebut the presumption of cross-applicability of prior art to those claims under 37 C.F.R. § 41.207(c). But that determination is premised on an erroneous construction that claim 1 and dependent claim 2 do not require a separate and ^ distinct third exhaust gas supplying step. Because we-reverse the Board’s construction of claim 1, we vacate its determination that claims [963]*9631 and 2 are unpatentable over Kinugasa '241, Kinugasa '024, and Binder '499, and remand for further proceedings under the proper claim construction.
The Board did not decide whether claims 1 and 2 of the '605 patent would have been obvious over Gandhi '135 and Surnilla '504, an issue raised by Gandhi. On remand, the Board will have an opportunity to address that issue.
V
We next consider whether the Board erred in construing claims 12-17 of the '901 application and in finding that the '901 application has an adequate written description of the “transitioning” limitation of those claims.
Sufficiency of written description is a question of fact, which we review for substantial evidence. Ariad Pharm., Inc. v. EliLilly & Co., 598 F.3d 1336, 1351 (Fed. Cir.2010) (en banc). Claims must be sufficiently supported by the written description of a patent, such that the disclosure “reasonably conveys to those skilled in the art that the inventor had possession of the claimed subject matter as of the filing date.” Id. “[W]hen a party challenges written description support for an interference count or the copied claim in an interference, the originating disclosure provides the meaning of the pertinent claim language.” Agilent, 567 F.3d at 1375. “When a party challenges a claim’s validity under § 102 or § 103, however, this court and the Board must interpret the claim in light of the specification in which it appears.” Id.
Beckmann argues that, when deciding the written description motion, the Board made a legal error by construing Gandhi’s claims 12-17 in light of the specification and prosecution history of Gandhi’s '901 application, contrary to Agilent. Beckmann contends that the proper construction requires the “transitioning” step to be separate and distinct. Under that construction, Beckmann argues, the '901 application lacks an adequate written description of the “transitioning” limitation because the application does not contain any express or inherent disclosure of a separate third exhaust gas supplying step controlled in terms of both exhaust gas composition and duration.
Gandhi responds that the Board’s construction is consistent with Beckmann’s '605 patent specification. Gandhi argues that the Board referred to extrinsic evidence only to support its understanding as to the plain meaning of the claim terms and to find written description support in the '901 application. Gandhi also responds that the Board properly declined to import limitations from the specification into the claims, and properly construed the claims as not requiring a separate and distinct transition step. Finally, Gandhi responds that substantial evidence supports the Board’s finding that Figure la and its related description in the '901 application provide an adequate written description of the “transitioning” limitation as the Board properly construed it.
Wé agree with Gandhi and the Board that claim 12 does not require a separate and distinct transitioning step between the lean and rich exhaust gas supplying steps, and that the plain language of the claim merely requires “transitioning” between lean and rich supplying steps and “completing” the transition from lean to rich. Here again, because the intrinsic record fully determines the proper construction, we review the Board’s construction de novo. See Microsoft, 789 F.3d at 1297.
Unlike claim 3 of the '901 application, which is identical to Beckmann’s claim 1, claim 12 of the '901 application only refers to “supplying” the lean and rich exhaust [964]*964gases as “supplying step[s],” but does not refer to “transitioning” between the lean and rich supplying steps or “completing” the transition as a step. J.A. 508 (“transitioning between said lean exhaust gas supplying step and said rich exhaust gas supplying step, wherein exhaust gas is supplied to said zoned catalyst system during said transition; and completing said transition from said lean exhaust gas supplying step to said rich exhaust gas supplying step ” (emphases added)). Moreover, claim 12 does not impose any limitation on the composition of exhaust gas being supplied during the transition to distinguish it from the lean and rich supplying steps; rather, the claim only requires that “exhaust gas [be] supplied ... during said transition.” Id. Accordingly, unlike Beckmann’s claim 1 and Gandhi’s claim 3, Gandhi’s claim 12 does not require a separate and distinct transitioning step. Nor does it require a separate and distinct third exhaust gas supplying step that is controlled in terms of exhaust gas composition and duration.
The specification of the '605 patent does not compel a different conclusion. On appeal, the parties do not dispute that claim 12, purportedly copied from Beckmann, should be construed in light of the specification of the '605 patent. See Agilent, 567 F.3d at 1375. The '605 patent repeatedly refers to the third intermediate exhaust gas supplying step as a “third method step.” '605 patent col. 4 1. 48-eol. 5 1. 67. The written description of the '605 patent uses the term “transition” only once, and that isolated usage does not define the term differently from its plain meaning. Id. col. 5 1. 2 (“during the transition to the second method step”). Accordingly, in light of the plain claim language and the '605 patent specification, we conclude that claim 12 of the '901 application does not require a separate and distinct transitioning step.
We also agree with Gandhi that substantial evidence supports the Board’s finding that the '901 application provides adequate written description support for the “transitioning” limitation of claims 12-17. Figure la of the '901 application illustrates the NOx output when the engine transitions from lean to rich operations. J.A. 495. The '901 application further describes “the short period in which the NOx trap transitions from lean to rich” and refers to Figure la as showing the NOx spikes during the “lean-rich transition.” J.A. 502. Indeed, on appeal, Beckmann only challenges the Board’s written description finding under its proposed narrower construction.
We therefore affirm the Board’s finding that the '901 application contains an adequate written description of the “transitioning” limitation of claims 12-17.
VI
Finally, we consider whether the Board erred in finding that the '901 application contains an adequate written description of the- “minimizing the oxygen content” limitation of claims 5, 7, 8, 10, 11, 16, and 17.
Beckmann argues that the Board erred by relying on paragraph 41 of the '901 application, which describes “removing oxygen storage capacity” of a lean NOx trap, to find support for the “minimizing” limitation. According to Beckmann, the Board conflated the.physical characteristics of a lean NOx trap, which is described by paragraph 41, with a method step that takes place “prior to rich conditions,” which is what the claims require. Beckmann argues that, reading paragraph 41 and the following paragraphs in "their entireties, one of ordinary skill would understand that the description of “removing oxygen storage capacity refers to how a lean NOx trap could be constructed, not how it could [965]*965be used during engine operation. Gandhi responds that substantial evidence supports the Board’s finding that the '901 application provides adequate written description support for the “minimizing” limitation, and that the Board did not err in not crediting Beckmann’s expert testimony on that issue, which did not specifically address paragraph 41.
We conclude that substantial evidence supports the Board’s written description finding regarding the “minimizing the oxygen content” limitation of claims 7, 10, and 16, but not as to claims 5, 8, 11, and 17. The claims at issue are all directed to “a method ... comprising” a list of actions, including “minimizing the oxygen content” of the NOx trap. But those claims differ in that claims 5, 8, 11, and 17 contain a “minimizing the oxygen content ... prior to a rich cycle” limitation, whereas claims 7, 10, and 16 recite that “the oxygen content of the lean NOx trap is minimized” without requiring that the minimization occur “prior to” a rich cycle. Notably, claims 8, 11, and 17 respectively depend from claims 7, 10, and 16, suggesting that the “prior to” language imposes a temporal requirement that “minimizing” take place “prior to” rich conditions as part of the claimed method.
The Board did not consider and analyze whether the '901 application provides an adequate written description of the “prior to” limitation. We therefore vacate its written description decision as to claims 5, 8,11, and 17.
We do agree with the Board, however, that paragraph 41 of the '901 application provides an adequate written description of the “minimizing the oxygen content” limitation of claim's 7, 10, and 16 because those claims do not require that “minimizing” occur before a rich cycle. Even if the disclosures in paragraph 41 of the '901 application are limited to designing and constructing an NOx trap to minimize its oxygen storage capacity, that description is sufficient to support the “minimizing” limitation of claims 7, 10, and 16. We therefore affirm the Board’s decision that the '901 application contains an adequate written description of the “minimizing the oxygen content” limitation of claims 7, 10, and 16.
Conclusion
We have considered the parties’ remaining arguments, but find them unpersuasive. For the foregoing reasons, we reverse the Board’s construction of claims 1 and 2 of the '605 patent, but we affirm the Board’s decision that claims 3 and 4 of the '605 patent would have been obvious over the cited references and that the '901 application contains an adequate written description of the “transitioning” limitation of claims 12-17 and the “minimizing the oxygen content” limitation of claims 7, 10, and 16. Additionally, we vacate the Board’s determination 'that claims 1 and 2 of the '605 patent are unpatentable over the cited references and its finding that the '901 application contains an adequate written description of.the “minimizing the oxygen content ... prior to a rich cycle” limitation of claims 5, 8, 11, and 17, and remand for further proceedings consistent with this opinion.
AFFIRMED IN PART, REVERSED IN PART, VACATED IN PART, AND REMANDED
Costs
No costs.