Lee Jiyun, Lee Jaehoon, Bang Hyeonsu, Yoon Tae Woong, Ko Jong Hwan, Kang Boseok
SKKU Advanced Institute of Nanotechnology (SAINT) and Department of Nano Science and Technology, Sungkyunkwan University, 2066 Seobu-ro, Jangan-gu, Suwon 16419, Republic of Korea.
Department of Electrical and Computer Engineering, Sungkyunkwan University, 2066 Seobu-ro, Jangan-gu, Suwon 16419, Republic of Korea.
Mater Horiz. 2025 Jul 14;12(14):5331-5341. doi: 10.1039/d5mh00348b.
The limitations of traditional von Neumann architectures have driven interest in organic mixed ionic-electronic conductors (OMIECs) for integrating memory and computation. Organic electrochemical synaptic transistors (OESTs) are particularly promising for emulating biological synaptic behaviors because they offer low power consumption, flexibility, and scalability. One-shot integrable electropolymerization (OSIEP) has emerged as a promising approach for fabricating OESTs owing to its simplicity and integrative capabilities. However, OSIEP-fabricated devices often exhibit inferior memory characteristics, largely due to suboptimal control of channel crystallinity-a key factor influencing memory retention. In this study, we addressed this challenge by fabricating poly(3,4-ethylenedioxythiphene) (PEDOT)-based OESTs using a mixed binary supporting electrolyte the OSIEP method. A binary system comprising tetrabutylammonium tetrafluoroborate (BF) and 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (TFSI) was adopted to balance crystallinity and ionic conductivity. PEDOT:Blend films achieved enhanced synaptic functionality by combining the high de-doping efficiency and charge transport of PEDOT:BF with the superior molecular orientation of PEDOT:TFSI. This synergistic approach significantly improved the long-term depression/potentiation characteristics and prolonged memory retention. PEDOT:Blend-based synaptic transistors achieved a recognition accuracy of 95.58% on the MNIST dataset, surpassing devices fabricated with single electrolytes. These findings highlight a scalable strategy for tuning the synaptic properties in OMIEC-based devices, thereby advancing their potential for neuromorphic computing applications.
传统冯·诺依曼架构的局限性激发了人们对用于集成存储和计算的有机混合离子-电子导体（OMIECs）的兴趣。有机电化学突触晶体管（OESTs）因其低功耗、灵活性和可扩展性，在模拟生物突触行为方面特别有前景。单次可集成电聚合（OSIEP）因其简单性和集成能力，已成为制造OESTs的一种有前景的方法。然而，OSIEP制造的器件通常表现出较差的存储特性，这主要是由于对沟道结晶度的控制不理想——这是影响记忆保持的关键因素。在本研究中，我们通过使用混合二元支持电解质的OSIEP方法制造基于聚（3,4-乙撑二氧噻吩）（PEDOT）的OESTs来应对这一挑战。采用由四丁基铵四氟硼酸盐（BF）和1-乙基-3-甲基咪唑双（三氟甲基磺酰）亚胺（TFSI）组成的二元体系来平衡结晶度和离子电导率。PEDOT:混合膜通过将PEDOT:BF的高去掺杂效率和电荷传输与PEDOT:TFSI的优异分子取向相结合，实现了增强的突触功能。这种协同方法显著改善了长期抑制/增强特性，并延长了记忆保持时间。基于PEDOT:混合膜的突触晶体管在MNIST数据集上实现了95.58%的识别准确率，超过了用单一电解质制造的器件。这些发现突出了一种可扩展的策略，用于调整基于OMIEC的器件中的突触特性，从而提升它们在神经形态计算应用中的潜力。
