An electrode in a lithium-ion cell is usually treated as one substance. You mix active material, binder, conductive additive and — in a solid-state cell — solid electrolyte, you coat it, you press it, and every point through that coating is supposed to look like every other point. A US patent application published on July 16, 2026 and assigned to TOYOTA JIDOSHA KABUSHIKI KAISHA takes the opposite position: it treats one of the two electrodes as a material whose recipe should deliberately change from its top face to its bottom face. The application, US20260204733A1, is titled BATTERY AND METHOD FOR MANUFACTURING BATTERY. It is a pending application — published, not granted — with six named inventors and a CPC spread that runs across separators (H01M 50/461, H01M 50/449), electrode composition (H01M 4/366, H01M 4/386, H01M 4/667) and cell architecture (H01M 10/0525, H01M 10/0585).

Start with the geometry, because the claimed architecture is more specific than the summary description suggests. Claim 1 is not a single sandwich. It builds outward symmetrically from one central member: a first current collector layer carrying a pair of first electrode layers, one on each of its two surfaces. On each of those sits a solid electrolyte layer, on each of those a second electrode layer, and on each of those an outer second current collector layer. Everything above the centre line is mirrored below it. That double-sided limitation — "disposed on one surface and the other surface" — is a requirement of the claim itself; the application's abstract describes the stack without it, so the abstract should be read as background description rather than as a statement of what is claimed.

The grading, and what it is trying to fix

The point of novelty sits inside the second electrode layer. That layer is a composite: it contains electrode active material and solid electrolyte together, which is standard practice in a bulk-type solid-state cell, where the electrolyte inside the electrode is the only thing carrying lithium ions from the separator to the active particles. What is not standard is that the ratio between those two ingredients is claimed to vary along the thickness direction.

wherein the second electrode layers each contain an electrode active material and a solid electrolyte, and a content ratio of the solid electrolyte is higher on the solid electrolyte layer side than on the second current collector layer side in a thickness direction.— BATTERY AND METHOD FOR MANUFACTURING BATTERY, US20260204733A1

Read that as a gradient with two ends doing two different jobs. The face of the electrode that touches the solid electrolyte layer is the ionic handoff point — lithium arrives there and has to spread into the electrode without piling up, so that face is electrolyte-rich. The face that touches the outer current collector is the electronic handoff point — electrons have to reach the metal foil, and solid electrolyte is an ionic conductor, not an electronic one, so more electrolyte there is dead weight and added resistance. A uniform mix has to compromise between the two. A graded layer does not have to.

Notably, claim 1 itself never says which electrode is which. It speaks only of "first" and "second" electrode layers, which leaves the independent claim agnostic as to polarity. The assignment arrives in dependent claim 3, which identifies the first current collector and first electrode layers as the cathode side and the second electrode layers — the graded ones — as the anode layers, with the outer collectors as anode current collectors. So the graded electrode is the anode, and it is an anode reading that lives at claim 3, not at claim 1.

Claim 4 narrows one step further, adding that the anode layers each contain silicon as the electrode active material. That is two levels of dependency down from the independent claim, so silicon is a specific embodiment rather than a requirement of the broadest claim — but it makes the engineering rationale legible. Silicon anodes store far more lithium per gram than graphite and swell substantially while doing it. In a liquid cell the electrolyte simply reflows around the expanding particles. In a solid-state cell nothing reflows; every volume change has to be absorbed by a rigid composite, and losing contact at an interface means losing capacity. Tuning how much solid electrolyte sits at each face of a silicon anode is a direct response to that constraint.

Building a layer that isn't uniform

A gradient is easy to draw and hard to coat. The application answers that with a discretised construction. Claim 2 provides that at least one second electrode layer includes a plurality of partial second electrode layers laminated in the thickness direction, with the electrolyte-richer partial layer nearer the solid electrolyte layer and the leaner one nearer the outer collector. In other words, the smooth gradient is approximated by a small stack of sub-layers of stepped composition — a manufacturable proxy for a continuous profile.

Claim 5 turns that into a process. It is a transfer-based method: press the laminate of the central collector and its pair of first electrode layers; transfer a solid electrolyte layer onto each first electrode layer from a carrier substrate; then transfer a first partial second electrode layer onto each solid electrolyte layer, and a second partial second electrode layer onto each of those, with the electrolyte content of the first partial layer higher than that of the second. Each composition is prepared and handled on its own substrate and laid down in sequence, which sidesteps the problem of trying to coat a compositional gradient wet-on-wet.

The application does not sit alone in Toyota's July 16 publications. The same day's records include a SECONDARY BATTERY application, a METHOD OF MANUFACTURING POWER STORAGE DEVICE AND POWER STORAGE DEVICE, a further POWER STORAGE DEVICE, a BATTERY PACK AND BATTERY PACK MANUFACTURING METHOD and a MEMBRANE ELECTRODE ASSEMBLY. Taken together the cluster spans cell chemistry, module assembly and process, with a recurring emphasis on how layers are made and joined rather than on new active materials alone.

Here is what the record actually says, stripped of the framing. Toyota has applied for coverage on a double-sided solid-state stack whose anode is compositionally stratified along its thickness, built from stacked partial layers of stepped electrolyte content and laid down by transfer. Whether the application issues, and in what form, is for prosecution to determine. What it discloses is a view of the electrode as a depth-varying structure with two distinct interfaces to serve — which is a different engineering problem from the one a single uniform slurry is designed to solve.