US20260231568A1 published on August 6, 2026, assigned to LG Energy Solution, Ltd. and naming six inventors. Twenty claims, two of them independent: claim 1 to an isotope battery, claim 17 to a method of fabricating one. The classification codes on the record fall in the H10H 20-series — H10H 20/8312, 20/813, 20/822, 20/832, 20/835, 20/8512 and 20/8515 — which is semiconductor light-emitting device territory rather than the H01M electrochemical cell range where this applicant's filings normally sit. That placement is consistent with a device built as a semiconductor junction rather than as a cell.

Claim 1 has three elements and defines the invention by arrangement and shape:

a radiation source positioned along a first side of the first conductivity type semiconductor layer such that the radiation source is positioned in a first direction relative to the first conductivity type semiconductor layer, the radiation source having an at least partial droplet shape— Isotope Battery and Method of Fabricating the Same, US20260231568A1

The rest of claim 1 places a second conductivity type semiconductor layer in the same first direction, such that the radiation source ends up between the two layers. The scope-bearing words are “an at least partial droplet shape.” The claim recites no radionuclide, no activity, no half-life and no composition for the source; it recites what the source looks like. A conventional isotope battery with a uniformly deposited source film falls outside the claim not because of what the source is made of but because a film is not a droplet.

Where the dependents branch

Claims 2 and 3 seat the source on a flat surface extending perpendicular to the first direction. From there claims 4 and 5 split in a way worth reading carefully: claim 4 requires a contact angle greater than 90 degrees, claim 5 requires less than 90 degrees, and both depend from claim 3. They cannot both describe one embodiment. This is alternative claiming, not an inconsistency — a source bead that beads up and one that wets and spreads are genuinely different structures, and the applicant has claimed each rather than committing to one. Anyone reading claim 4 and claim 5 as contradictory has misread the dependency structure.

A separate branch runs through claim 6, which forms a concave pit in the first layer along the surface and seats the source at least partly inside it. Claim 7 then allows part of the source to protrude out of the pit, and claim 8 allows the protruding part to extend along the adjacent surface. Read in sequence, claims 6 through 8 walk from fully recessed to substantially proud, which covers the range of outcomes a dispensing process would actually produce against an etched target.

Claims 11 and 12 add further layers. Claim 11 adds a second conductivity type layer on the opposite side of the first layer, making a symmetric stack. Claim 12 adds an additional first conductivity type layer beyond the second layer. Claim 12 also contains a drafting defect that survives into the published text: it recites “positioned along a surface of (the second conductivity type semiconductor layer on an opposite side…” with an opening parenthesis that is never closed anywhere in the claim. The limitation is readable, but the punctuation is unbalanced as published.

Claims 13 and 14 introduce the one variant that changes the conversion mechanism: a photon generating layer between the radiation source and the first layer, with claim 14 requiring its interface with the source to follow an ellipsoidal upper surface. That is an indirect conversion path. Claims 15 and 16 multiply the source into discrete regions spaced along the side, each with the droplet shape, arranged in claim 16 as nodes of a grid.

Claims 9 and 10 handle two remaining geometric points. Claim 9 requires the first conductivity type layer to have a substantially constant thickness measured orthogonally to the surface, which forecloses embodiments that thin the layer beneath the source. Claim 10 requires the source to present at least a partially ellipsoidal surface — a shape characterisation that overlaps with the droplet language of claim 1 without being identical to it, since a droplet seated in the pit of claim 6 need not be ellipsoidal above the surface.

The method claim

Claim 17 recites three steps: distributing a radiation source slurry as a plurality of droplets spaced along the first surface; curing the distributed droplets; and forming the second conductivity type layer on the cured source and the first surface. Its dependents supply the process window. Claim 18 sets droplet diameter at discharge from 10 µm to 10 mm. Claim 20 sets slurry viscosity from 0.5 to 2000 cP at 25 °C. Claim 19 requires the ejected droplets to form a free surface along their total surface area before becoming seated — a limitation that distinguishes jetting from contact dispensing, and one of the few places the claim set constrains the equipment rather than the result.

Note that claim 17 does not incorporate claim 1 by reference. It is a standalone method independent, so its scope is not limited by claim 1's requirement that the cured source have an at least partial droplet shape in the finished device — although in practice the process it recites is what produces that shape. The two independents cover the same subject from opposite ends: one the artefact, one the act.

Context: this record published alongside forty-two other applications from the same applicant on August 6, nearly all of them lithium-ion — among them intumescent inter-cell fire protection, a cylindrical cell crimp specification and a lithium-rich manganese cathode. This is a pending application. Its claims are as filed and published, not as allowed.