Suppression of leakage in small-molecule memristors via fluoropolymer-assisted electric field modulation
Abstract The development of neuromorphic hardware systems is essential to meet the growing demands of artificial intelligence and internet of things technologies. Organic memristors (OMs), owing to their mechanical flexibility, low cost, biocompatibility, and tunable molecular properties, have emerged as promising candidates for synaptic elements in crossbar array architectures when constructing neuromorphic systems. Although vacuum-deposited OMs (V-OMs) offer high device uniformity and reliability—crucial for real-world applications—their high charge transport characteristics inherently lead to elevated leakage currents and crosstalk due to uncontrolled electric field distribution, thereby limiting system scalability and reliability. In this study, we propose a polymer-assisted field-confinement strategy to overcome these limitations by patterning a chemically compatible fluorinated polymer onto the active layer of V-OMs. The patterned fluoropolymer effectively confines the electric field to desired regions while suppressing it elsewhere, significantly reducing leakage current and enhancing electrical performance. The optimized devices exhibited over an order of magnitude reduction in high-resistance state current and an improved on/off ratio from 10 2 to 10 3 . Endurance and retention were validated through 300-cycle and 5000 s tests, respectively. Furthermore, stable pulse switching, along with multilevel conductance modulation under varying compliance currents, confirmed reliable and non-volatile synaptic weight updates. These results highlight the potential of this field-confinement strategy for realizing reliable, scalable, and high-density neuromorphic systems based on crossbar array structures.
Authors
- Hea‐Lim Park (ORCID: https://orcid.org/0000-0001-8370-7068)
- Jun-Seok Ro
Institutions
- Seoul National University of Science and Technology (KR)
Publication Details
- Journal
- Flexible and Printed Electronics
- Published
- 2025-09-01
- DOI
- https://doi.org/10.1088/2058-8585/ae01b8
- Citations
- 1
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.73
Funders
- National Research Foundation of Korea