Due to the aggressive scaling down of logic semiconductors, the difficulty of semiconductor component processes has increased. As the structure of components becomes more complex, the time and cost of processes and simulations have risen. Machine learning is now being used to analyze the electrical characteristics data of semiconductor components and apply the trained machine learning to next-generation semiconductor development. Machine learning trained on process data and simulation results can quickly and accurately predict which electrical characteristics change significantly when the component’s structure changes and which parameters have a significant impact on the electrical characteristic changes. This paper presents suitable machine learning models for analyzing and predicting the electrical characteristics (on-current ( I_on ), off-current ( I_off ), threshold voltage ( V_th ), subthreshold swing (SS), and drain induced barrier lowering (DIBL)) and statistical distribution (mean and standard deviation of the electrical characteristics) resulting from geometrical variability (sheet thickness ( T_wire ), sheet diameter ( D_wire ), oxide thickness ( T_ox ), gate length ( L_g ), spacer length ( L_sp ), gate metal work-function (WF)) in nanosheet field-effect transistor (NSFET), which are a next-generation logic device. Machine learning models, including regulation-based models (Ridge, LASSO) and tree-based models (decision tree (DT), random forest (RF), extreme gradient boost (XGBoost), and light gradient boost machine (LGBM)), are trained on technology computer-aided design (TCAD) simulation data. The LGBM more accurately predicts the electrical characteristics and statistical distribution of the NSFET than the other models. Additionally, we analyze the effect of geometrical variability on the NSFET based on feature importance.

Self-powered wireless sensor systems have emerged as an important topic for condition monitoring in nuclear power plants. However, commercial wireless sensor systems still cannot be fully self-sustainable due to the high power consumption caused by excessive signal processing in a mini-electronic computing system. In this sense, it is essential not only to integrate the sensor system with energy-harvesting devices but also to develop simple data processing methods for low power schemes. In this paper, we report a patch-type vibration visualization (PVV) sensor system based on the triboelectric effect and a visualization technique for self-sustainable operation. The PVV sensor system composed of a polyethylene terephthalate (PET)/Al/LCD screen directly converts the triboelectric signal into an informative black pattern on the LCD screen without excessive signal processing, enabling extremely low power operation. In addition, a proposed image processing method reconverts the black patterns to frequency and acceleration values through a remote-control camera. With these simple signal-to-pattern conversion and pattern-to-data reconversion techniques, a vibration visualization sensor network has successfully been demonstrated.