The research team led by Professor Hak-Yeol Bae (Division of Electronic Engineering) conducting studies for the commercialization of next‑generation neuromorphic computing analyzed the causes of performance degradation of next‑generation ferroelectric memory and neuromorphic semiconductor devices under radiation environments. The team also identified key reliability factors to consider when designing semiconductor devices for radiation environments.

The related research was carried out jointly by the research team of Professor Hak‑Yeol Bae at Jeonbuk National University (JBNU), the research team of Professor Tae‑Wan Kim at the University of Seoul, and Qbeam Solution, a startup from the Korea Atomic Energy Research Institute (CEO Bong‑Gi Jeong). The research paper was published in the latest issue of ACS Applied Electronic Materials (paper title: Neutron-Induced Defect-Driven Degradation of Ferroelectric Memory and Neuromorphic Synapses in a-In2Se3 FeS-FETs).

This study is notable for systematically linking and analyzing how defect‑related changes caused by neutron irradiation in two‑dimensional ferroelectric‑semiconductor‑based synaptic transistors affect electrical characteristics, memory operation, and neuromorphic synapse performance. In particular, the team confirmed that the observed performance degradation after neutron irradiation is not a simple reduction in current, but can lead to reduced charge transport due to defect‑related trap states and increased uncertainty in synaptic weights.

Ferroelectric semiconductors can retain a polarized state even when external power is turned off, making them promising candidates for next‑generation nonvolatile memory and neuromorphic computing devices. Neuromorphic computing is a next‑generation technology designed to perform memory and computation concurrently within a single device, similar to synapses in the human brain. It is regarded as an alternative capable of overcoming the limitations of conventional computing architectures, such as the von Neumann bottleneck.

However, in environments such as space and aircraft platforms, high‑energy neutrons can create defects inside semiconductors. Neutrons can penetrate deeply into materials, and the defects produced during this process can degrade the reliability of semiconductor devices.

To identify the causes of these issues, the research team performed neutron irradiation experiments. Energy‑dispersive X‑ray spectroscopy (EDS) analysis after neutron irradiation showed changes in the Se/In composition ratio of the thin film. Electrical analyses indicated an increase in acceptor‑like trap density, and these defect‑related trap states were interpreted to interfere with polarization‑controlled charge transport.

In practice, the device drain current decreased by approximately 88.95% after neutron irradiation, and the memory‑window‑to‑sweep‑range ratio (MW/SR), which represents memory characteristics, deteriorated by about 48.20%. In contrast, ferroelectric polarization switching characteristics were preserved, suggesting that the observed performance degradation is closely related to charge trapping and reduced transport caused by defect‑related trap states induced by neutron irradiation rather than complete loss of polarization.

The research team also confirmed that this degradation affects neuromorphic synapse operation. After neutron irradiation, the nonlinearity of synaptic weight updates increased and the dynamic range decreased, leading to reduced inference accuracy of artificial neural networks. This indicates that radiation‑induced defects can affect not only the electrical degradation of individual devices but also the reliability of AI computations at the system level.

The significance of this study lies in experimentally identifying degradation factors that must be considered when designing devices for radiation environments. By confirming that defect‑related trap states induced by neutron irradiation can sequentially affect polarization‑based charge transport, memory performance, and synaptic weight stability, the study provides foundational evidence needed for reliability evaluation of next‑generation semiconductor devices operating in radiation environments.

Professor Hak‑Yeol Bae stated, "For next‑generation AI semiconductors to operate reliably in extreme environments such as space, it is important to systematically understand the effects of radiation‑induced semiconductor defects beyond simply improving performance." He added that this study is significant because it analyzed the degradation pathways of neutron‑irradiated, two‑dimensional ferroelectric‑semiconductor‑based memory and neuromorphic devices and identified key physical factors to consider in future design and reliability evaluation of devices for radiation environments.

The study was supported by the Ministry of Science and ICT's Nano‑Materials Technology Development Project (Materials Global Young Connect), the vdW Materials and Process Technology Development Project for Ultra‑High‑Integration Semiconductors, and the JBNU RISE Project Group.