[Revisit 4] Project Janus Revisited: The Two-Faced Gate - A Rare-Earth-Oxide Memory Gate on the Angelica GaN HEMT that Forgets on Purpose: Trap Physics, Inverse Design of the Host Mixture, Three Dials, a Physical Reservoir, Applications and the Audit

Project Janus Revisited (revised edition, November 2026) of the Angelica Paper Series. A europium-doped rare-earth oxide (Eu:RE₂O₃) in the gate of an AlGaN/GaN HEMT is designed as a volatile memory whose forgetting time is chosen by chemistry, spacer thickness and temperature, from microseconds to about a month (Sc₂O₃ ceiling ≈ 34 d at 25 °C). The analysis is done on a single mobility-gap energy scale with measured band offsets. The decisive quantity is the lift, meaning how far the empty Eu²⁺ level sits above the electron layer. Elastic writing is possible only when the lift is below 2^(2/3)−1 ≈ 0.59 of the trap depth (Project Charon). On the standard Al₀.₂₈GaN barrier the lift is about 1.04 eV and writing is impossible. An Al-rich top barrier (Al ≈ 0.43–0.71) opens a window. Because the electron layer is a semiconductor, each cell writes inside a band of gate fields rather than above a threshold, and the field picks which trap depth is addressed. As a result, a gentle, long signal is kept and a strong, brief one is soon forgotten. In simulations, a 50-cell physical reservoir matches a tuned echo-state network on Lorenz forecasting and does better under read noise, while the network wins on NARMA10 and memory capacity. Hole-storage (Pr/Tb) and ferroelectric branches are left to companion papers. Every claim is tagged IMPLEMENTED, TESTABLE or CONCEPT, and a prioritised list of lab tests is given. The paper comes in three languages (en/zh-TW/ja) with primers. Text and figures are CC BY 4.0, and code is Apache-2.0. This is published as a defensive publication: anyone is free to use it, and it may not be patented by others. Keywords: GaN HEMT; europium; rare-earth oxide gate; charge-trap memory; volatile memory; tunable retention; band offset; lift rule; trap-assisted tunnelling; AlGaN barrier engineering; physical reservoir computing; neuromorphic hardware; spaced learning; Sc2O3; Y2O3; defensive publication; Angelica Paper Series

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-15
DOI
https://doi.org/10.5281/zenodo.24250080
Primary Topic
GaN-based semiconductor devices and materials
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

[Revisit 4] Project Janus Revisited: The Two-Faced Gate - A Rare-Earth-Oxide Memory Gate on the Angelica GaN HEMT that Forgets on Purpose: Trap Physics, Inverse Design of the Host Mixture, Three Dials, a Physical Reservoir, Applications and the Audit

Tung Ning Liu
Zenodo (CERN European Organization for Nuclear Research)
GaN-based semiconductor devices and materials
preprint

[Revisit 4] Project Janus Revisited: The Two-Faced Gate - A Rare-Earth-Oxide Memory Gate on the Angelica GaN HEMT that Forgets on Purpose: Trap Physics, Inverse Design of the Host Mixture, Three Dials, a Physical Reservoir, Applications and the Audit

Tung Ning Liu
preprint en

Abstract

Project Janus Revisited (revised edition, November 2026) of the Angelica Paper Series. A europium-doped rare-earth oxide (Eu:RE₂O₃) in the gate of an AlGaN/GaN HEMT is designed as a volatile memory whose forgetting time is chosen by chemistry, spacer thickness and temperature, from microseconds to about a month (Sc₂O₃ ceiling ≈ 34 d at 25 °C). The analysis is done on a single mobility-gap energy scale with measured band offsets. The decisive quantity is the lift, meaning how far the empty Eu²⁺ level sits above the electron layer. Elastic writing is possible only when the lift is below 2^(2/3)−1 ≈ 0.59 of the trap depth (Project Charon). On the standard Al₀.₂₈GaN barrier the lift is about 1.04 eV and writing is impossible. An Al-rich top barrier (Al ≈ 0.43–0.71) opens a window. Because the electron layer is a semiconductor, each cell writes inside a band of gate fields rather than above a threshold, and the field picks which trap depth is addressed. As a result, a gentle, long signal is kept and a strong, brief one is soon forgotten. In simulations, a 50-cell physical reservoir matches a tuned echo-state network on Lorenz forecasting and does better under read noise, while the network wins on NARMA10 and memory capacity. Hole-storage (Pr/Tb) and ferroelectric branches are left to companion papers. Every claim is tagged IMPLEMENTED, TESTABLE or CONCEPT, and a prioritised list of lab tests is given. The paper comes in three languages (en/zh-TW/ja) with primers. Text and figures are CC BY 4.0, and code is Apache-2.0. This is published as a defensive publication: anyone is free to use it, and it may not be patented by others. Keywords: GaN HEMT; europium; rare-earth oxide gate; charge-trap memory; volatile memory; tunable retention; band offset; lift rule; trap-assisted tunnelling; AlGaN barrier engineering; physical reservoir computing; neuromorphic hardware; spaced learning; Sc2O3; Y2O3; defensive publication; Angelica Paper Series

Zenodo (CERN European Organization for Nuclear Research)
GaN-based semiconductor devices and materials
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon