Synaptic plasticity in Co/Nb:STO memristive devices: the role of oxygen vacancies
Walter Quiñonez, Anouk Goossens, Diego Rubi, Tamalika Banerjee, María José Sánchez · 2025
Read Synaptic plasticity in Co/Nb:STO memristive devices: the role of oxygen vacancies by Walter Quiñonez, Anouk Goossens, Diego Rubi, Tamalika Banerjee, María José Sánchez (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-11965.txt.
Introduction
A chip that remembers like a synapse is the neuromorphic dream, and Walter Quinonez and four co-authors traced the memory in one device to wandering oxygen atoms, in a 2025 paper. Their device is a cobalt contact on niobium-doped strontium titanate, forming a Schottky barrier whose height sets the resistance.
Pulses moved the resistance in two tempos: a power-law drift while reading, the signature of short-term plasticity, and stepwise climbs with repeated pulses, the signature of long-term memory. The team's model reproduces both from one mechanism, oxygen vacancies migrating and reshaping the barrier. Matching experiment to model matters because a synapse chip must be designed, not just discovered; knowing which atoms carry the memory tells engineers which knobs will tune it.
Catalog record
| CREATOR | Walter Quiñonez, Anouk Goossens, Diego Rubi, Tamalika Banerjee, María José Sánchez |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 27 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-11965
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Synaptic plasticity in Co/Nb:STO memristive devices: the role of oxygen vacancies Walter Quiñonez, Anouk Goossens, Diego Rubi, Tamalika Banerjee, María José Sánchez Abstract Neuromorphic computing aims to develop energy-efficient devices that mimic biological synapses. One promising approach involves memristive devices, that can dynamically adjust their electrical resistance in response to stimuli, similar to synaptic weight changes in the brain. However, a key challenge is understanding and controlling the coexistence of different types of synaptic plasticity, like short-term and long-term plasticity. In this work, we show that plasticity behaviors in Co/Nb:STO Schottky memristors originate from oxygen vacancy electromigration, which modulates the Schottky barrier and enables both short-term and long-term plasticity. Our experiments reveal that resistance changes follow a power-law during reading (short-term plasticity) and increases stepwise with successive pulses (long-term memory retention). These behaviors are successfully reproduced by our model that demonstrates the correlation between oxygen vacancy distribution and Schottky barrier modulation. Our findings highlight these memristors as promising candidates for neuromorphic applications. I Introduction The drive to overcome the limitations imposed by the von Neumann bottleneck has spurred the development of advanced materials and devices that transcend conventional silicon technologies Najmaei et al. 2022 ; Park et al. 2023 . Neuromorphic computing has emerged as a promising paradigm designed to achieve more efficient and brain-inspired information processing Upadhyay et al. 2019 ; Kendall and Kumar 2020 ; Zhang et al. 2020 . The principal requisites for neuromorphic devices can be broadly categorized into energy efficiency, compact integration, and brain-like functionalities, which are further divided into neuronal and synaptic functionalities to emulate different aspects of brain-like processing. Among the various technologies underpinning neuromorphic computing, memristors Valov and Kozicki 2013 ; Kumar et al. 2022 have emerged as key enablers for hardware implementation of such architectures Li et al. 2018 ; Mehonic et al. 2020 ; Sebastian et al. 2020 ; Yao et al. 2020 ; Sun, Chen, and Yan 2021 ; Zhou et al. 2022 . Memristors are metal/insulator/metal capacitor-like structures that exhibit resistive switching (RS), where electrical resistance changes reversibly in response to applied electrical stimuli Sawa 2008 ; Yang et al. 2008 ; Waser and Aono 2010 ; Borghetti et al. 2010 ; Panda and Tseng 2014 ; Wang et al. 2020 . Functional oxides are particularly well suited materials to built memristive devices due to their highly tunable physical and electrical properties, responsiveness to external stimuli and promising compatibility with modern CMOS integration Park et al. 2023 . Consequently, oxides-based memristors Waser and Aono 2010 ; Lee et al. 2011 ; Breuer et al. 2016 ; Kumar et al. 2022 ; Yongyue Xiao and Ye 2023 ; Patil et al. 2023 are valued due to their highly tunable physical and electrical properties, low power consumption, potential scalability and their ability to exhibit non-volatile and/or volatile RS. In non-volatile (NV) RS, resistive states persist even after the electrical stimulus is removed. In particular, memristors that exhibit NV RS are promising candidates for future memory devices due to their simple structure, fast switching time, high density integration, low power consumption, retention times and 3D stacking possibility Waser et al. 2009 ; Waser and Aono 2010 . In contrast, volatile RS (VRS) exhibits resistive states that spontaneously relax once the electrical stimulus is turned off Zhuo et al. 2013 ; Zhou et al. 2022 . NV memristors can effectively mimic biological long-term plasticity, enabling the construction of artificial neural networks using crossbar arrays Perez et al. 2019 ; Xia and Yang 2019 . VRS, on the other hand, can emulate short-term plasticity, which plays a crucial role in neural network dynamics and is particularly well-suited for reservoir computing systems Ielmini and Wong 2018 . A variety of synaptic plasticity mechanisms have been implemented in hardware, including long-term and short-term potentiation and depression Ielmini and Ambrogio 2019 ; Zhuo et al. 2013 ; Zhou et al. 2022 as well as spike-timing-dependent plasticity (STDP), among others. These characteristics naturally facilitate "in-memory" computation Wu et al. 2023 , mirroring the adaptive synaptic strengths of neural networks, which must be dynamically updated as learning occurs. The modulation of the Schottky barrier formed at the metal-insulator interface due to the dynamics of oxygen vacancies (OV) is one of the most widely accepted mechanism for interfacial RS in several oxide-based memristors Sawa 2008 ; Rozenberg et al. 2010 ; Ielmini and Waser 2016 ; Valov 2017 ; Panda and Tseng 2014 . In particular, metal/Nb:SrTiO 3 (STO) Schottky junctions have been extensively investigated and several works show that the emergence of RS is governed by the quality of the metal/STO interface Pan et al. 2014 ; Mikheev et al. 2014 . We have demonstrated in earlier works Goossens et al. 2023 ; Goossens, Das, and Banerjee 2018 ; Goossens and Banerjee 2023 interface Schottky memristors using Co electrodes on Nb:STO. Such interfaces exploit the non-linear variation of the dielectric permittivity in Nb:STO with electric field and exhibit multi-level switching with large ON/OFF ratios and minimal device-to-device and cycle-to-cycle variations Goossens et al. 2023 ; Goossens, Das, and Banerjee 2018 . In addition, cross-sectional scanning transmission electron microscopy (STEM) combined with energy-dispersive X-ray spectroscopy (EDS) Goossens et al. 2023 , reveal the presence of an homogeneous oxygen-deficient layer at the interface which allows for the trapping and detrapping of carriers. By performing an electrostatic analysis Goossens and Banerjee 2023 , a DR length of approximately 15 nm was estimated for Co/Nb:STO devices with a doping concentration of 0.1 wt \% . We have also studied how memristive characteristics such as ON/OFF ratio, stochasticity and nonlinearity are influenced by varying the doping of the Nb:STO semiconductor Goossens and Banerjee 2023 , opening prospects in applications such as data encryption and random number generation. In this study, we fabricated such two-terminal Co/Nb:STO memristors, with the primary goal of exploring their capability to mimic synaptic plasticity. We specifically focus on the potentiation of resistive states under electrical pulsing schemes with varying parameters. Our experiments revealed a combination of NVRS and VRS behaviors, emulating an intertwined effect of LTP and STP. Furthermore, we observed that these processes are governed by power laws with distinct exponents. Based on these findings, we propose a theoretical model that accounts for OV electromigration, demonstrating how the movement of OV modulates the Schottky barrier over time, thereby reproducing the electrical response observed in the experiments. This work provides valuable insights into the mechanisms underlying synaptic behavior in Co/Nb:STO memristors, attributing it to OV electromigration rather than conventional models of RS based on charge trapping and detrapping, where defects are treated as fixed traps within the material. II Methods For the fabrication of the devices (see Fig. 1 (a)), Nb:STO (001) substrates with doping concentration of 0.1 wt \% from Crystec was used. The substrate was chemically treated with buffered hydrofluoric acid (BHF) and subsequently annealed at 960 {}^{\circ}\text{C} in an O 2 flow of 300 ccmin -1 . Electron beam lithography and electron beam evaporation were used to fabricate circular Co electrodes (20 nm thick) with a radius of 2 \mu m. The electrodes were electrically isolated by an insulating AlO x layer to minimize device crosstalk. To prevent oxidation and protect the contacts, a 100 nm