EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Junction‐Programmed Excitability Control in a Thin‐Film Bipolar Neuromorphic Transistor

Emad Iranmanesh, Mingzhe Song, Yuanhao Xia, Charalabos Doumanidis, Hang Zhou

Advanced Electronic Materials · 22.11.2026

DOI: 10.1002/aelm.70656 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Advanced Electronic Materials, online erschienen am 22.11.2026. Verfasst von Emad Iranmanesh und 4 weiteren Autor:innen. Die Arbeit nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National University of Ireland.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Control of neuronal excitability is central to adaptive neuromorphic hardware, yet device‐defined states are often represented as synaptic quantities before being translated into neuronal activity. Junction‐defined excitability is demonstrated here in a thin‐film bipolar neuromorphic transistor, where distinct emitter–base (E–B) conditions directly regulate the membrane‐driving current and spike generation under fixed leaky‐integrate‐and‐fire parameters, without an intermediate synaptic‐weight representation. Quantitative transport analysis links the excitability states to relative effective E–B barrier shifts and corresponding changes in membrane drive. Principal‐state, multilevel, and four‐device measurements demonstrate systematic and reproducible modulation of firing rate, inter‐spike interval, and temporal firing statistics, together with operational persistence over the neuronal measurement interval. A device‐calibrated spiking neural network constrained by the measured firing rate and inter‐spike‐interval variability transfers the experimental states into stochastic neuronal dynamics. The resulting states produce strongly differentiated early‐inference and hidden‐layer spiking while maintaining approximately 96.2% final classification accuracy. These results establish direct excitability‐to‐spike transfer through a semiconductor junction and identify the E–B junction as a controllable neuronal excitability interface for adaptive event‐driven neuromorphic hardware.

Förderung

  • National University of Ireland

Zitieren

APA

Iranmanesh, E., Song, M., Xia, Y., Doumanidis, C., & Zhou, H. (2026). Junction‐Programmed Excitability Control in a Thin‐Film Bipolar Neuromorphic Transistor. Advanced Electronic Materials. https://doi.org/10.1002/aelm.70656

Vancouver

Iranmanesh E, Song M, Xia Y, Doumanidis C, Zhou H. Junction‐Programmed Excitability Control in a Thin‐Film Bipolar Neuromorphic Transistor. Advanced Electronic Materials. 2026. doi:10.1002/aelm.70656

BibTeX
@article{Iranmanesh2026,
  author = {Iranmanesh, Emad and Song, Mingzhe and Xia, Yuanhao and Doumanidis, Charalabos and Zhou, Hang},
  title = {Junction‐Programmed Excitability Control in a Thin‐Film Bipolar Neuromorphic Transistor},
  journal = {Advanced Electronic Materials},
  year = {2026},
  doi = {10.1002/aelm.70656}
}
Kostenlos starten →