Answers are grounded in this paper and cite their sources.
Second edition (November 2026). It replaces the first edition (June 2026), which placed one graphene sheet at the Gd₂O₃/GaN interface of the Angelica GaN-on-Si stack and reported stress relief, a denser 2DEG and a 1.9–4.8 meV Rashba spin–orbit coupling driving a field-free spin Hall current. Rechecked on the engine of Angelica Revisited, all three are withdrawn. (1) The claims. The Rashba estimate scaled the coupling that heavy transition-metal dichalcogenides lend to graphene; GaN has only light atoms. With the first-principles field term (5 µeV per V/nm of the field in graphene's van der Waals gap) and the largest computed proximity coupling on GaN (65 µeV), the coupling is at most about 135 µeV anywhere in the stack, 14–36 times below the claim and far below the disorder broadening of transferred graphene; a spin Hall effect from it is negligible. A sheet transferred after growth changes the GaN stress by at most 1.7 MPa (claimed: 15). The 2DEG gain is a threshold shift: graphene on the barrier is the gate. (2) Where a sheet can sit. Between the memory trap and the channel, a contacted sheet keeps the memory's threshold shift but cuts the gate's coupling to the channel to 15–48 %. Buried under the GaN it must survive growth at 1050 °C. As the gate it does not write the Eu level of the Janus memory (Charon's lift rule), even when a negative gate dopes the sheet. (3) The light window. Two stores of the series are written by light (Ce in Styx, Pr in Janus). Transfer-matrix optics of the whole stack show that behind graphene the trap layer receives 0.89–1.20 of the incident light from 270 to 535 nm, about as much as with no gate; 50 nm of ITO falls to 0.26 at 270 nm and 5 nm of Pt gives 0.48–0.65. (4) The 2 µm GaN on Si is an optical cavity at visible wavelengths: with a laser and a ±2.5 % thickness spread, the delivered light ranges 0.21–2.5 across a wafer; an LED of 15 nm bandwidth narrows it to 1.19–1.21. (5) Writing by light costs about 0.8–80 J/cm² per bit, 6–8 orders of magnitude above an electrical update, so these stores suit rare writes. Projected light addresses cells without wear; flood light with a gate veto wears the oxide within a few to tens of writes. The writing light also empties the Eu trace, so writing the long-term store forgets the short-term one. Files: paper in English, Traditional Chinese and Japanese (the English version governs); a primer in each language; code and sources (code_and_sources.zip) reproducing every number and figure. Computational throughout; nothing is measured. Part of the Angelica Paper Series. Licence: text and figures CC BY 4.0; code Apache-2.0. Defensive publication: disclosed as prior art, free for anyone to use and not to be patented by others. This edition replaces the first edition's licences (CC BY-NC-ND 4.0; BUSL 1.1).
No takes yet. Share an insight, caveat, or question.
Tung Ning Liu (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: