Journal Article

FZJ-2022-03250

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Filamentary TaO x /HfO 2 ReRAM Devices for Neural Networks Training with Analog In‐Memory Computing

Stecconi, T. (Corresponding author) ; Guido, R. ; Berchialla, L. ; La Porta, A. ; Weiss, J. ; Popoff, Y. ; Halter, M. ; Sousa, M. ; Horst, F. ; Dávila, D. ; Drechsler, U. ; Dittmann, R.FZJ* ; Offrein, B. J. ; Bragaglia, V.

2022
Wiley-VCH Verlag GmbH & Co. KG Weinheim

This record in other databases:    

Please use a persistent id in citations: http://hdl.handle.net/2128/32064  doi:10.1002/aelm.202200448

Abstract: The in-memory computing paradigm aims at overcoming the intrinsic inefficiencies of Von-Neumann computers by reducing the data-transport per arithmetic operation. Crossbar arrays of multilevel memristive devices enable efficient calculations of matrix-vector-multiplications, an operation extensively called on in artificial intelligence (AI) tasks. Resistive random-access memories (ReRAMs) are promising candidate devices for such applications. However, they generally exhibit large stochasticity and device-to-device variability. The integration of a sub-stoichiometric metal-oxide within the ReRAM stack can improve the resistive switching graduality and stochasticity. To this purpose, a conductive TaOx layer is developed and stacked on HfO2 between TiN electrodes, to create a complementary metal-oxide-semiconductor-compatible ReRAM structure. This device shows accumulative conductance updates in both directions, as required for training neural networks. Moreover, by reducing the TaOx thickness and by increasing its resistivity, the device resistive states increase, as required for reduced power consumption. An electric field-driven TaOx oxidation/reduction is responsible for the ReRAM switching. To demonstrate the potential of the optimized TaOx/HfO2 devices, the training of a fully-connected neural network on the Modified National Institute of Standards and Technology database dataset is simulated and benchmarked against a full precision digital implementation.

---

Contributing Institute(s):

1. Elektronische Materialien (PGI-7)
2. JARA-FIT (JARA-FIT)

Research Program(s):

1. 5233 - Memristive Materials and Devices (POF4-523) (POF4-523)
2. MANIC - Materials for Neuromorphic Circuits (861153) (861153)
3. DFG project 167917811 - SFB 917: Resistiv schaltende Chalkogenide für zukünftige Elektronikanwendungen: Struktur, Kinetik und Bauelementskalierung "Nanoswitches" (167917811) (167917811)
4. BMBF-16ME0398K - Verbundprojekt: Neuro-inspirierte Technologien der künstlichen Intelligenz für die Elektronik der Zukunft - NEUROTEC II - (BMBF-16ME0398K) (BMBF-16ME0398K)
5. BMBF-16ME0404 - Verbundprojekt: Neuro-inspirierte Technologien der künstlichen Intelligenz für die Elektronik der Zukunft - NEUROTEC II - (BMBF-16ME0404) (BMBF-16ME0404)
6. BMBF-03ZU1106AB - NeuroSys: "Memristor Crossbar Architekturen (Projekt A) - B" (BMBF-03ZU1106AB) (BMBF-03ZU1106AB)
7. ACA - Advanced Computing Architectures (SO-092) (SO-092)

Appears in the scientific report 2022

---

Database coverage:
; ; ; Clarivate Analytics Master Journal List ; Current Contents - Physical, Chemical and Earth Sciences ; DEAL Wiley ; Essential Science Indicators ; IF >= 5 ; JCR ; SCOPUS ; Science Citation Index Expanded ; Web of Science Core Collection

---