Non‐Perturbative Oxygen Vacancy Mapping via Independently‐Contacted, Reciprocally‐Switching Double‐Layer ECRAM for In‐Memory Computing
The electrochemical control of ion migration is fundamental to the development of oxide memristors and neuromorphic devices. Among them, oxygen-ion-based electrochemical random-access memory (ECRAM) offers low variability and high training accuracy, making it suitable for analog in-memory computing. However, conventional vertically-stacked ECRAMs hinder real-time visualization of ion migration paths, limiting further device optimization. Here, an independently-contacted, reciprocally-switching double-layer ECRAM (IRIS-ECRAM) is introduced that enables simultaneous measurement of conductance changes in both the channel and reservoir layers. This design allows direct mapping of oxygen vacancy migration within the device, revealing opposing switching behavior between the two layers and identifying a new role of the electrolyte as a temporary reservoir. Furthermore, array-level operation, confirming the scalability of the proposed design, is demonstrated. This platform is universally applicable to various ECRAM structures and provides a powerful tool for understanding ionic dynamics and guiding structural optimization based on oxygen vacancy mapping.
Authors
- Hyunjeong Kwak (ORCID: https://orcid.org/0009-0008-8814-1823)
- Seungmin Han (ORCID: https://orcid.org/0009-0005-4495-3406)
- Seungkun Kim (ORCID: https://orcid.org/0009-0004-1378-0699)
- Jung‐Ho Lee (ORCID: https://orcid.org/0000-0002-6731-3111)
- Jeonghoon Son
- Young‐Joon Lee (ORCID: https://orcid.org/0000-0002-1735-3385)
- Changwook Jeong (ORCID: https://orcid.org/0000-0003-3962-7153)
- Hyeongjun Jang (ORCID: https://orcid.org/0009-0003-1455-7395)
- Beomjoo Ham (ORCID: https://orcid.org/0009-0003-2090-8067)
- Jinho Byun
Institutions
- Pohang University of Science and Technology (KR)
- Ulsan National Institute of Science and Technology (KR)
Publication Details
- Journal
- Small
- Published
- 2025-11-24
- DOI
- https://doi.org/10.1002/smll.202508271
- Primary Topic
- Advanced Memory and Neural Computing
- Type
- article
- Field-Weighted Citation Impact
- 0.00